René's URL Explorer Experiment


Title: GitHub - turned123/NLP-Interview-Notes: 本项目是作者们根据个人面试和经验总结出的自然语言处理(NLP)面试准备的学习笔记与资料，该资料目前包含 自然语言处理各领域的 面试题积累。 · GitHub

Open Graph Title: GitHub - turned123/NLP-Interview-Notes: 本项目是作者们根据个人面试和经验总结出的自然语言处理(NLP)面试准备的学习笔记与资料，该资料目前包含 自然语言处理各领域的 面试题积累。

X Title: GitHub - turned123/NLP-Interview-Notes: 本项目是作者们根据个人面试和经验总结出的自然语言处理(NLP)面试准备的学习笔记与资料，该资料目前包含 自然语言处理各领域的 面试题积累。

Description: 本项目是作者们根据个人面试和经验总结出的自然语言处理(NLP)面试准备的学习笔记与资料，该资料目前包含 自然语言处理各领域的 面试题积累。 - turned123/NLP-Interview-Notes

Open Graph Description: 本项目是作者们根据个人面试和经验总结出的自然语言处理(NLP)面试准备的学习笔记与资料，该资料目前包含 自然语言处理各领域的 面试题积累。 - turned123/NLP-Interview-Notes

X Description: 本项目是作者们根据个人面试和经验总结出的自然语言处理(NLP)面试准备的学习笔记与资料，该资料目前包含 自然语言处理各领域的 面试题积累。 - turned123/NLP-Interview-Notes

Opengraph URL: https://github.com/turned123/NLP-Interview-Notes

X: @github

direct link

Domain: github.com

route-pattern/:user_id/:repository
route-controllerfiles
route-actiondisambiguate
fetch-noncev2:bcbccac2-6fe2-aaec-f6dc-d4c9930db056
current-catalog-service-hashf3abb0cc802f3d7b95fc8762b94bdcb13bf39634c40c357301c4aa1d67a256fb
request-idD51C:1C3DD7:1B6DE1F:248D6EC:6A52CE7A
html-safe-nonce5db87f15c5b2e690dd4fda02d4581bd8fe930fce0df11929dc3b968b591b8195
visitor-payloadeyJyZWZlcnJlciI6IiIsInJlcXVlc3RfaWQiOiJENTFDOjFDM0RENzoxQjZERTFGOjI0OEQ2RUM6NkE1MkNFN0EiLCJ2aXNpdG9yX2lkIjoiOTE2MTg4OTY1MDUxNTgyNDI1MCIsInJlZ2lvbl9lZGdlIjoiaWFkIiwicmVnaW9uX3JlbmRlciI6ImlhZCJ9
visitor-hmacb253d244cddd231ec990801218287eb3b5e1291f3ee665dd4f4c49f6757ef5f2
hovercard-subject-tagrepository:377779664
github-keyboard-shortcutsrepository,copilot
google-site-verificationApib7-x98H0j5cPqHWwSMm6dNU4GmODRoqxLiDzdx9I
octolytics-urlhttps://collector.github.com/github/collect
analytics-location//
fb:app_id1401488693436528
apple-itunes-appapp-id=1477376905, app-argument=https://github.com/turned123/NLP-Interview-Notes
twitter:imagehttps://opengraph.githubassets.com/81002b7759387aa87e656b82cb7bbdfa9a0f5dc47396b038a67276e6eec21ef7/turned123/NLP-Interview-Notes
twitter:cardsummary_large_image
og:imagehttps://opengraph.githubassets.com/81002b7759387aa87e656b82cb7bbdfa9a0f5dc47396b038a67276e6eec21ef7/turned123/NLP-Interview-Notes
og:image:alt本项目是作者们根据个人面试和经验总结出的自然语言处理(NLP)面试准备的学习笔记与资料，该资料目前包含 自然语言处理各领域的 面试题积累。 - turned123/NLP-Interview-Notes
og:image:width1200
og:image:height600
og:site_nameGitHub
og:typeobject
hostnamegithub.com
expected-hostnamegithub.com
Noneb9a586c06a05a7a86fc7e3f4dbd03e42f6869085879aa184aa6369456dbd50fb
turbo-cache-controlno-cache
go-importgithub.com/turned123/NLP-Interview-Notes git https://github.com/turned123/NLP-Interview-Notes.git
octolytics-dimension-user_id19400627
octolytics-dimension-user_loginturned123
octolytics-dimension-repository_id377779664
octolytics-dimension-repository_nwoturned123/NLP-Interview-Notes
octolytics-dimension-repository_publictrue
octolytics-dimension-repository_is_forktrue
octolytics-dimension-repository_parent_id377670349
octolytics-dimension-repository_parent_nwoleoymr/NLP-Interview-Notes
octolytics-dimension-repository_network_root_id341198747
octolytics-dimension-repository_network_root_nwokm1994/NLP-Interview-Notes
turbo-body-classeslogged-out env-production page-responsive
disable-turbofalse
browser-stats-urlhttps://api.github.com/_private/browser/stats
browser-errors-urlhttps://api.github.com/_private/browser/errors
release07a982c1d40157c619b364352b704c3ce66bb332
ui-targetfull
theme-color#1e2327
color-schemelight dark

Links:

Skip to contenthttps://github.com/turned123/NLP-Interview-Notes#start-of-content
https://github.com/
Sign in https://github.com/login?return_to=https%3A%2F%2Fgithub.com%2Fturned123%2FNLP-Interview-Notes
GitHub CopilotWrite better code with AIhttps://github.com/features/copilot
GitHub Copilot appDirect agents from issue to mergehttps://github.com/features/ai/github-app
MCP RegistryNewIntegrate external toolshttps://github.com/mcp
ActionsAutomate any workflowhttps://github.com/features/actions
CodespacesInstant dev environmentshttps://github.com/features/codespaces
IssuesPlan and track workhttps://github.com/features/issues
Code ReviewManage code changeshttps://github.com/features/code-review
GitHub Advanced SecurityFind and fix vulnerabilitieshttps://github.com/security/advanced-security
Code securitySecure your code as you buildhttps://github.com/security/advanced-security/code-security
Secret protectionStop leaks before they starthttps://github.com/security/advanced-security/secret-protection
Why GitHubhttps://github.com/why-github
Documentationhttps://docs.github.com
Bloghttps://github.blog
Changeloghttps://github.blog/changelog
Marketplacehttps://github.com/marketplace
View all featureshttps://github.com/features
Enterpriseshttps://github.com/enterprise
Small and medium teamshttps://github.com/team
Startupshttps://github.com/enterprise/startups
Nonprofitshttps://github.com/solutions/industry/nonprofits
App Modernizationhttps://github.com/solutions/use-case/app-modernization
DevSecOpshttps://github.com/solutions/use-case/devsecops
DevOpshttps://github.com/solutions/use-case/devops
CI/CDhttps://github.com/solutions/use-case/ci-cd
View all use caseshttps://github.com/solutions/use-case
Healthcarehttps://github.com/solutions/industry/healthcare
Financial serviceshttps://github.com/solutions/industry/financial-services
Manufacturinghttps://github.com/solutions/industry/manufacturing
Governmenthttps://github.com/solutions/industry/government
View all industrieshttps://github.com/solutions/industry
View all solutionshttps://github.com/solutions
AIhttps://github.com/resources/articles?topic=ai
Software Developmenthttps://github.com/resources/articles?topic=software-development
DevOpshttps://github.com/resources/articles?topic=devops
Securityhttps://github.com/resources/articles?topic=security
View all topicshttps://github.com/resources/articles
Customer storieshttps://github.com/customer-stories
Events & webinarshttps://github.com/resources/events
Ebooks & reportshttps://github.com/resources/whitepapers
Business insightshttps://github.com/solutions/executive-insights
GitHub Skillshttps://skills.github.com
Documentationhttps://docs.github.com
Customer supporthttps://support.github.com
Community forumhttps://github.com/orgs/community/discussions
Trust centerhttps://github.com/trust-center
Partnershttps://github.com/partners
View all resourceshttps://github.com/resources
GitHub SponsorsFund open source developershttps://github.com/open-source/sponsors
Security Labhttps://securitylab.github.com
Maintainer Communityhttps://maintainers.github.com
Acceleratorhttps://github.com/open-source/accelerator
GitHub Starshttps://stars.github.com
Archive Programhttps://archiveprogram.github.com
Topicshttps://github.com/topics
Trendinghttps://github.com/trending
Collectionshttps://github.com/collections
Enterprise platformAI-powered developer platformhttps://github.com/enterprise
GitHub Advanced SecurityEnterprise-grade security featureshttps://github.com/security/advanced-security
Copilot for BusinessEnterprise-grade AI featureshttps://github.com/features/copilot/copilot-business
Premium SupportEnterprise-grade 24/7 supporthttps://github.com/enterprise/premium-support
Pricinghttps://github.com/pricing
Search syntax tipshttps://docs.github.com/search-github/github-code-search/understanding-github-code-search-syntax
documentationhttps://docs.github.com/search-github/github-code-search/understanding-github-code-search-syntax
Sign in https://github.com/login?return_to=https%3A%2F%2Fgithub.com%2Fturned123%2FNLP-Interview-Notes
Sign up https://github.com/signup?ref_cta=Sign+up&ref_loc=header+logged+out&ref_page=%2F%3Cuser-name%3E%2F%3Crepo-name%3E&source=header-repo&source_repo=turned123%2FNLP-Interview-Notes
Reloadhttps://github.com/turned123/NLP-Interview-Notes
Reloadhttps://github.com/turned123/NLP-Interview-Notes
Reloadhttps://github.com/turned123/NLP-Interview-Notes
turned123 https://github.com/turned123
NLP-Interview-Noteshttps://github.com/turned123/NLP-Interview-Notes
leoymr/NLP-Interview-Noteshttps://github.com/leoymr/NLP-Interview-Notes
Notifications https://github.com/login?return_to=%2Fturned123%2FNLP-Interview-Notes
Fork 0 https://github.com/login?return_to=%2Fturned123%2FNLP-Interview-Notes
Star 0 https://github.com/login?return_to=%2Fturned123%2FNLP-Interview-Notes
Code https://github.com/turned123/NLP-Interview-Notes
Pull requests 0 https://github.com/turned123/NLP-Interview-Notes/pulls
Actions https://github.com/turned123/NLP-Interview-Notes/actions
Projects https://github.com/turned123/NLP-Interview-Notes/projects
Security and quality 0 https://github.com/turned123/NLP-Interview-Notes/security
Insights https://github.com/turned123/NLP-Interview-Notes/pulse
Code https://github.com/turned123/NLP-Interview-Notes
Pull requests https://github.com/turned123/NLP-Interview-Notes/pulls
Actions https://github.com/turned123/NLP-Interview-Notes/actions
Projects https://github.com/turned123/NLP-Interview-Notes/projects
Security and quality https://github.com/turned123/NLP-Interview-Notes/security
Insights https://github.com/turned123/NLP-Interview-Notes/pulse
https://github.com/turned123/NLP-Interview-Notes
Brancheshttps://github.com/turned123/NLP-Interview-Notes/branches
Tagshttps://github.com/turned123/NLP-Interview-Notes/tags
https://github.com/turned123/NLP-Interview-Notes/branches
https://github.com/turned123/NLP-Interview-Notes/tags
19 Commitshttps://github.com/turned123/NLP-Interview-Notes/commits/main/
https://github.com/turned123/NLP-Interview-Notes/commits/main/
BasicAlgorithmhttps://github.com/turned123/NLP-Interview-Notes/tree/main/BasicAlgorithm
BasicAlgorithmhttps://github.com/turned123/NLP-Interview-Notes/tree/main/BasicAlgorithm
DeepLearningAlgorithmhttps://github.com/turned123/NLP-Interview-Notes/tree/main/DeepLearningAlgorithm
DeepLearningAlgorithmhttps://github.com/turned123/NLP-Interview-Notes/tree/main/DeepLearningAlgorithm
MachineLearningAlgorithmhttps://github.com/turned123/NLP-Interview-Notes/tree/main/MachineLearningAlgorithm
MachineLearningAlgorithmhttps://github.com/turned123/NLP-Interview-Notes/tree/main/MachineLearningAlgorithm
NLPinterviewhttps://github.com/turned123/NLP-Interview-Notes/tree/main/NLPinterview
NLPinterviewhttps://github.com/turned123/NLP-Interview-Notes/tree/main/NLPinterview
Tensorflowhttps://github.com/turned123/NLP-Interview-Notes/tree/main/Tensorflow
Tensorflowhttps://github.com/turned123/NLP-Interview-Notes/tree/main/Tensorflow
Trickhttps://github.com/turned123/NLP-Interview-Notes/tree/main/Trick
Trickhttps://github.com/turned123/NLP-Interview-Notes/tree/main/Trick
imghttps://github.com/turned123/NLP-Interview-Notes/tree/main/img
imghttps://github.com/turned123/NLP-Interview-Notes/tree/main/img
pythonhttps://github.com/turned123/NLP-Interview-Notes/tree/main/python
pythonhttps://github.com/turned123/NLP-Interview-Notes/tree/main/python
resourcehttps://github.com/turned123/NLP-Interview-Notes/tree/main/resource
resourcehttps://github.com/turned123/NLP-Interview-Notes/tree/main/resource
xmindhttps://github.com/turned123/NLP-Interview-Notes/tree/main/xmind
xmindhttps://github.com/turned123/NLP-Interview-Notes/tree/main/xmind
zoomhttps://github.com/turned123/NLP-Interview-Notes/tree/main/zoom
zoomhttps://github.com/turned123/NLP-Interview-Notes/tree/main/zoom
模板https://github.com/turned123/NLP-Interview-Notes/tree/main/%E6%A8%A1%E6%9D%BF
模板https://github.com/turned123/NLP-Interview-Notes/tree/main/%E6%A8%A1%E6%9D%BF
README.en.mdhttps://github.com/turned123/NLP-Interview-Notes/blob/main/README.en.md
README.en.mdhttps://github.com/turned123/NLP-Interview-Notes/blob/main/README.en.md
README.mdhttps://github.com/turned123/NLP-Interview-Notes/blob/main/README.md
README.mdhttps://github.com/turned123/NLP-Interview-Notes/blob/main/README.md
README.pnghttps://github.com/turned123/NLP-Interview-Notes/blob/main/README.png
README.pnghttps://github.com/turned123/NLP-Interview-Notes/blob/main/README.png
READMEhttps://github.com/turned123/NLP-Interview-Notes
https://github.com/turned123/NLP-Interview-Notes#关于-nlp百问百答
https://github.com/km1994/NLP-Interview-Noteshttps://github.com/km1994/NLP-Interview-Notes
https://github.com/km1994/nlp_paper_studyhttps://github.com/km1994/nlp_paper_study
https://github.com/turned123/NLP-Interview-Notes/blob/main/img/20210523220743.png
https://github.com/turned123/NLP-Interview-Notes/blob/main/img/%E5%BE%AE%E4%BF%A1%E6%88%AA%E5%9B%BE_20210301212242.png
https://github.com/turned123/NLP-Interview-Notes/blob/main/img/%E5%BE%AE%E4%BF%A1%E6%88%AA%E5%9B%BE_20210212153059.png
https://github.com/turned123/NLP-Interview-Notes#介绍
https://github.com/turned123/NLP-Interview-Notes#目录架构
【关于 NLP】百问百答https://github.com/turned123/NLP-Interview-Notes#%E5%85%B3%E4%BA%8E-nlp%E7%99%BE%E9%97%AE%E7%99%BE%E7%AD%94
介绍https://github.com/turned123/NLP-Interview-Notes#%E4%BB%8B%E7%BB%8D
目录架构https://github.com/turned123/NLP-Interview-Notes#%E7%9B%AE%E5%BD%95%E6%9E%B6%E6%9E%84
内容框架https://github.com/turned123/NLP-Interview-Notes#%E5%86%85%E5%AE%B9%E6%A1%86%E6%9E%B6
一、【关于 基础算法篇】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#%E4%B8%80%E5%85%B3%E4%BA%8E-%E5%9F%BA%E7%A1%80%E7%AE%97%E6%B3%95%E7%AF%87%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
二、【关于 机器学习算法篇】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#%E4%BA%8C%E5%85%B3%E4%BA%8E-%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E7%AE%97%E6%B3%95%E7%AF%87%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
三、【关于 深度学习算法篇】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#%E4%B8%89%E5%85%B3%E4%BA%8E-%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AE%97%E6%B3%95%E7%AF%87%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
四、【关于 NLP 学习算法】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#%E5%9B%9B%E5%85%B3%E4%BA%8E-nlp-%E5%AD%A6%E4%B9%A0%E7%AE%97%E6%B3%95%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.1 【关于 信息抽取】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#41-%E5%85%B3%E4%BA%8E-%E4%BF%A1%E6%81%AF%E6%8A%BD%E5%8F%96%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.1.1 【关于 命名实体识别】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#411-%E5%85%B3%E4%BA%8E-%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.1.2 【关于 关系抽取】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#412-%E5%85%B3%E4%BA%8E-%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.1.3 【关于 事件抽取】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#413-%E5%85%B3%E4%BA%8E-%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.2 【关于 NLP 预训练算法】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#42-%E5%85%B3%E4%BA%8E-nlp-%E9%A2%84%E8%AE%AD%E7%BB%83%E7%AE%97%E6%B3%95%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.3 【关于 文本分类】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#43-%E5%85%B3%E4%BA%8E-%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.4 【关于 文本匹配】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#44-%E5%85%B3%E4%BA%8E-%E6%96%87%E6%9C%AC%E5%8C%B9%E9%85%8D%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.5 【关于 问答系统】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#45-%E5%85%B3%E4%BA%8E-%E9%97%AE%E7%AD%94%E7%B3%BB%E7%BB%9F%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.5.1 【关于 FAQ 检索式问答系统】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#451-%E5%85%B3%E4%BA%8E-faq-%E6%A3%80%E7%B4%A2%E5%BC%8F%E9%97%AE%E7%AD%94%E7%B3%BB%E7%BB%9F-%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.5.2 【关于 问答系统工具篇】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#452-%E5%85%B3%E4%BA%8E-%E9%97%AE%E7%AD%94%E7%B3%BB%E7%BB%9F%E5%B7%A5%E5%85%B7%E7%AF%87-%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.6 【关于 对话系统】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#46-%E5%85%B3%E4%BA%8E-%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.7 【关于 知识图谱】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#47-%E5%85%B3%E4%BA%8E-%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.7.1 【关于 知识图谱】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#471-%E5%85%B3%E4%BA%8E-%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1-%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.7.2 【关于 KBQA】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#472-%E5%85%B3%E4%BA%8E-kbqa%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.7.3 【关于 Neo4j】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#473-%E5%85%B3%E4%BA%8E-neo4j%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.8 【关于 文本摘要】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#48-%E5%85%B3%E4%BA%8E-%E6%96%87%E6%9C%AC%E6%91%98%E8%A6%81-%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.9 【关于 知识表示学习】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#49-%E5%85%B3%E4%BA%8E-%E7%9F%A5%E8%AF%86%E8%A1%A8%E7%A4%BA%E5%AD%A6%E4%B9%A0%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
4.10 【关于 数据挖掘】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#410--%E5%85%B3%E4%BA%8E-%E6%95%B0%E6%8D%AE%E6%8C%96%E6%8E%98%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
五、【关于 NLP 技巧】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#%E4%BA%94%E5%85%B3%E4%BA%8E-nlp-%E6%8A%80%E5%B7%A7%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
5.1 【关于 少样本问题】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#51-%E5%85%B3%E4%BA%8E-%E5%B0%91%E6%A0%B7%E6%9C%AC%E9%97%AE%E9%A2%98%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
5.2 【关于 脏数据】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#52-%E5%85%B3%E4%BA%8E-%E8%84%8F%E6%95%B0%E6%8D%AE%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
5.3 【关于 炼丹炉】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#53-%E5%85%B3%E4%BA%8E-%E7%82%BC%E4%B8%B9%E7%82%89%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
5.4 【关于 早停法 EarlyStopping 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#54-%E5%85%B3%E4%BA%8E-%E6%97%A9%E5%81%9C%E6%B3%95-earlystopping-%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
5.5 【关于 标签平滑法 LabelSmoothing 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#55-%E5%85%B3%E4%BA%8E-%E6%A0%87%E7%AD%BE%E5%B9%B3%E6%BB%91%E6%B3%95-labelsmoothing-%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
六、【关于 Python 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#%E5%85%AD%E5%85%B3%E4%BA%8E-python-%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
七、【关于 Tensorflow 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes#%E4%B8%83%E5%85%B3%E4%BA%8E-tensorflow-%E9%82%A3%E4%BA%9B%E4%BD%A0%E4%B8%8D%E7%9F%A5%E9%81%93%E7%9A%84%E4%BA%8B
https://github.com/turned123/NLP-Interview-Notes#内容框架
【关于 基础算法篇】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm
https://github.com/turned123/NLP-Interview-Notes#一关于-基础算法篇那些你不知道的事
【关于 过拟合和欠拟合】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E8%BF%87%E6%8B%9F%E5%90%88%E5%92%8C%E6%AC%A0%E6%8B%9F%E5%90%88.md
一、过拟合和欠拟合 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E8%BF%87%E6%8B%9F%E5%90%88%E5%92%8C%E6%AC%A0%E6%8B%9F%E5%90%88.md#%E4%B8%80%E8%BF%87%E6%8B%9F%E5%90%88%E5%92%8C%E6%AC%A0%E6%8B%9F%E5%90%88-%E6%98%AF%E4%BB%80%E4%B9%88
二、过拟合/高方差（overfiting / high variance）篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E8%BF%87%E6%8B%9F%E5%90%88%E5%92%8C%E6%AC%A0%E6%8B%9F%E5%90%88.md#%E4%BA%8C%E8%BF%87%E6%8B%9F%E5%90%88%E9%AB%98%E6%96%B9%E5%B7%AEoverfiting--high-variance%E7%AF%87
2.1 过拟合是什么及检验方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E8%BF%87%E6%8B%9F%E5%90%88%E5%92%8C%E6%AC%A0%E6%8B%9F%E5%90%88.md#21-%E8%BF%87%E6%8B%9F%E5%90%88%E6%98%AF%E4%BB%80%E4%B9%88%E5%8F%8A%E6%A3%80%E9%AA%8C%E6%96%B9%E6%B3%95
2.2 导致过拟合的原因是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E8%BF%87%E6%8B%9F%E5%90%88%E5%92%8C%E6%AC%A0%E6%8B%9F%E5%90%88.md#22-%E5%AF%BC%E8%87%B4%E8%BF%87%E6%8B%9F%E5%90%88%E7%9A%84%E5%8E%9F%E5%9B%A0%E6%98%AF%E4%BB%80%E4%B9%88
2.3 过拟合的解决方法是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E8%BF%87%E6%8B%9F%E5%90%88%E5%92%8C%E6%AC%A0%E6%8B%9F%E5%90%88.md#23-%E8%BF%87%E6%8B%9F%E5%90%88%E7%9A%84%E8%A7%A3%E5%86%B3%E6%96%B9%E6%B3%95%E6%98%AF%E4%BB%80%E4%B9%88
三、欠拟合/高偏差（underfiting / high bias）篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E8%BF%87%E6%8B%9F%E5%90%88%E5%92%8C%E6%AC%A0%E6%8B%9F%E5%90%88.md#%E4%B8%89%E6%AC%A0%E6%8B%9F%E5%90%88%E9%AB%98%E5%81%8F%E5%B7%AEunderfiting--high-bias%E7%AF%87
3.1 欠拟合是什么及检验方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E8%BF%87%E6%8B%9F%E5%90%88%E5%92%8C%E6%AC%A0%E6%8B%9F%E5%90%88.md#31-%E6%AC%A0%E6%8B%9F%E5%90%88%E6%98%AF%E4%BB%80%E4%B9%88%E5%8F%8A%E6%A3%80%E9%AA%8C%E6%96%B9%E6%B3%95
3.2 导致欠拟合的原因是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E8%BF%87%E6%8B%9F%E5%90%88%E5%92%8C%E6%AC%A0%E6%8B%9F%E5%90%88.md#32-%E5%AF%BC%E8%87%B4%E6%AC%A0%E6%8B%9F%E5%90%88%E7%9A%84%E5%8E%9F%E5%9B%A0%E6%98%AF%E4%BB%80%E4%B9%88
3.3 过拟合的解决方法是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E8%BF%87%E6%8B%9F%E5%90%88%E5%92%8C%E6%AC%A0%E6%8B%9F%E5%90%88.md#33-%E8%BF%87%E6%8B%9F%E5%90%88%E7%9A%84%E8%A7%A3%E5%86%B3%E6%96%B9%E6%B3%95%E6%98%AF%E4%BB%80%E4%B9%88
【关于 BatchNorm vs LayerNorm】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 独立同分布（independent and identically distributed）与白化https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#11-%E7%8B%AC%E7%AB%8B%E5%90%8C%E5%88%86%E5%B8%83independent-and-identically-distributed%E4%B8%8E%E7%99%BD%E5%8C%96
1.2 （ Internal Covariate Shift，ICS）https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#12--internal-covariate-shiftics
1.3 ICS问题带来的后果是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#13-ics%E9%97%AE%E9%A2%98%E5%B8%A6%E6%9D%A5%E7%9A%84%E5%90%8E%E6%9E%9C%E6%98%AF%E4%BB%80%E4%B9%88
二、Normalization 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#%E4%BA%8Cnormalization-%E7%AF%87
2.1 Normalization 的通用框架与基本思想https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#21-normalization-%E7%9A%84%E9%80%9A%E7%94%A8%E6%A1%86%E6%9E%B6%E4%B8%8E%E5%9F%BA%E6%9C%AC%E6%80%9D%E6%83%B3
三、Batch Normalization 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#%E4%B8%89batch-normalization-%E7%AF%87
3.1 Batch Normalization（纵向规范化）是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#31-batch-normalization%E7%BA%B5%E5%90%91%E8%A7%84%E8%8C%83%E5%8C%96%E6%98%AF%E4%BB%80%E4%B9%88
3.2 Batch Normalization（纵向规范化）存在什么问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#32-batch-normalization%E7%BA%B5%E5%90%91%E8%A7%84%E8%8C%83%E5%8C%96%E5%AD%98%E5%9C%A8%E4%BB%80%E4%B9%88%E9%97%AE%E9%A2%98
3.3 Batch Normalization（纵向规范化）适用的场景是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#33-batch-normalization%E7%BA%B5%E5%90%91%E8%A7%84%E8%8C%83%E5%8C%96%E9%80%82%E7%94%A8%E7%9A%84%E5%9C%BA%E6%99%AF%E6%98%AF%E4%BB%80%E4%B9%88
3.4 BatchNorm 存在什么问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#34-batchnorm-%E5%AD%98%E5%9C%A8%E4%BB%80%E4%B9%88%E9%97%AE%E9%A2%98
四、Layer Normalization（横向规范化） 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#%E5%9B%9Blayer-normalization%E6%A8%AA%E5%90%91%E8%A7%84%E8%8C%83%E5%8C%96-%E7%AF%87
4.1 Layer Normalization（横向规范化）是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#41-layer-normalization%E6%A8%AA%E5%90%91%E8%A7%84%E8%8C%83%E5%8C%96%E6%98%AF%E4%BB%80%E4%B9%88
4.2 Layer Normalization（横向规范化）有什么用？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#42-layer-normalization%E6%A8%AA%E5%90%91%E8%A7%84%E8%8C%83%E5%8C%96%E6%9C%89%E4%BB%80%E4%B9%88%E7%94%A8
五、BN vs LN 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#%E4%BA%94bn-vs-ln-%E7%AF%87
六、主流 Normalization 方法为什么有效？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/BatchNormVsLayerNorm.md#%E5%85%AD%E4%B8%BB%E6%B5%81-normalization-%E6%96%B9%E6%B3%95%E4%B8%BA%E4%BB%80%E4%B9%88%E6%9C%89%E6%95%88
【关于 激活函数】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 为什么要有激活函数？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#11-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E6%9C%89%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0
二、激活函数介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#%E4%BA%8C%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1 sigmoid 函数篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#21-sigmoid-%E5%87%BD%E6%95%B0%E7%AF%87
2.1.1 什么是 sigmoid 函数？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#211-%E4%BB%80%E4%B9%88%E6%98%AF-sigmoid-%E5%87%BD%E6%95%B0
2.1.2 为什么选 sigmoid 函数 作为激活函数？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#212-%E4%B8%BA%E4%BB%80%E4%B9%88%E9%80%89-sigmoid-%E5%87%BD%E6%95%B0-%E4%BD%9C%E4%B8%BA%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0
2.1.3 sigmoid 函数 有什么缺点？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#213-sigmoid-%E5%87%BD%E6%95%B0-%E6%9C%89%E4%BB%80%E4%B9%88%E7%BC%BA%E7%82%B9
2.2 tanh 函数篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#22-tanh-%E5%87%BD%E6%95%B0%E7%AF%87
2.2.1 什么是 tanh 函数？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#221-%E4%BB%80%E4%B9%88%E6%98%AF-tanh-%E5%87%BD%E6%95%B0
2.2.2 为什么选 tanh 函数 作为激活函数？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#222-%E4%B8%BA%E4%BB%80%E4%B9%88%E9%80%89-tanh-%E5%87%BD%E6%95%B0-%E4%BD%9C%E4%B8%BA%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0
2.2.3 tanh 函数 有什么缺点？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#223-tanh-%E5%87%BD%E6%95%B0-%E6%9C%89%E4%BB%80%E4%B9%88%E7%BC%BA%E7%82%B9
2.3 relu 函数篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#23-relu-%E5%87%BD%E6%95%B0%E7%AF%87
2.3.1 什么是 relu 函数？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#231-%E4%BB%80%E4%B9%88%E6%98%AF-relu-%E5%87%BD%E6%95%B0
2.3.2 为什么选 relu 函数 作为激活函数？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#232-%E4%B8%BA%E4%BB%80%E4%B9%88%E9%80%89-relu-%E5%87%BD%E6%95%B0-%E4%BD%9C%E4%B8%BA%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0
2.3.3 relu 函数 有什么缺点？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#233-relu-%E5%87%BD%E6%95%B0-%E6%9C%89%E4%BB%80%E4%B9%88%E7%BC%BA%E7%82%B9
三、激活函数选择篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0.md#%E4%B8%89%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0%E9%80%89%E6%8B%A9%E7%AF%87
【关于 正则化】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md
一、L0，L1，L2正则化 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#%E4%B8%80l0l1l2%E6%AD%A3%E5%88%99%E5%8C%96-%E7%AF%87
1.1 正则化 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#11-%E6%AD%A3%E5%88%99%E5%8C%96-%E6%98%AF%E4%BB%80%E4%B9%88
1.2 什么是 L0 正则化 ？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#12-%E4%BB%80%E4%B9%88%E6%98%AF-l0-%E6%AD%A3%E5%88%99%E5%8C%96-
1.3 什么是 L1 （稀疏规则算子 Lasso regularization）正则化 ？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#13-%E4%BB%80%E4%B9%88%E6%98%AF-l1-%E7%A8%80%E7%96%8F%E8%A7%84%E5%88%99%E7%AE%97%E5%AD%90-lasso-regularization%E6%AD%A3%E5%88%99%E5%8C%96-
1.4 什么是 L2 正则化（岭回归 Ridge Regression 或者 权重衰减 Weight Decay）正则化 ？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#14-%E4%BB%80%E4%B9%88%E6%98%AF-l2-%E6%AD%A3%E5%88%99%E5%8C%96%E5%B2%AD%E5%9B%9E%E5%BD%92-ridge-regression-%E6%88%96%E8%80%85-%E6%9D%83%E9%87%8D%E8%A1%B0%E5%87%8F-weight-decay%E6%AD%A3%E5%88%99%E5%8C%96-
二、对比篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#%E4%BA%8C%E5%AF%B9%E6%AF%94%E7%AF%87
2.1 什么是结构风险最小化？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#21-%E4%BB%80%E4%B9%88%E6%98%AF%E7%BB%93%E6%9E%84%E9%A3%8E%E9%99%A9%E6%9C%80%E5%B0%8F%E5%8C%96
2.2 从结构风险最小化的角度理解L1和L2正则化https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#22-%E4%BB%8E%E7%BB%93%E6%9E%84%E9%A3%8E%E9%99%A9%E6%9C%80%E5%B0%8F%E5%8C%96%E7%9A%84%E8%A7%92%E5%BA%A6%E7%90%86%E8%A7%A3l1%E5%92%8Cl2%E6%AD%A3%E5%88%99%E5%8C%96
2.3 L1 vs L2https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#23-l1-vs-l2
三、dropout 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#%E4%B8%89dropout-%E7%AF%87
3.1 什么是 dropout？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#31-%E4%BB%80%E4%B9%88%E6%98%AF-dropout
3.2 dropout 在训练和测试过程中如何操作？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#32-dropout-%E5%9C%A8%E8%AE%AD%E7%BB%83%E5%92%8C%E6%B5%8B%E8%AF%95%E8%BF%87%E7%A8%8B%E4%B8%AD%E5%A6%82%E4%BD%95%E6%93%8D%E4%BD%9C
3.3 dropout 如何防止过拟合?https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.mdBasicAlgorithm/%E6%AD%A3%E5%88%99%E5%8C%96.md#33-dropout-%E5%A6%82%E4%BD%95%E9%98%B2%E6%AD%A2%E8%BF%87%E6%8B%9F%E5%90%88
【关于 优化算法及函数】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 为什么需要 优化函数？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#11-%E4%B8%BA%E4%BB%80%E4%B9%88%E9%9C%80%E8%A6%81-%E4%BC%98%E5%8C%96%E5%87%BD%E6%95%B0
1.2 优化函数的基本框架是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#12-%E4%BC%98%E5%8C%96%E5%87%BD%E6%95%B0%E7%9A%84%E5%9F%BA%E6%9C%AC%E6%A1%86%E6%9E%B6%E6%98%AF%E4%BB%80%E4%B9%88
二、优化函数介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#%E4%BA%8C%E4%BC%98%E5%8C%96%E5%87%BD%E6%95%B0%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1 梯度下降法是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#21-%E6%A2%AF%E5%BA%A6%E4%B8%8B%E9%99%8D%E6%B3%95%E6%98%AF%E4%BB%80%E4%B9%88
2.2 随机梯度下降法是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#22-%E9%9A%8F%E6%9C%BA%E6%A2%AF%E5%BA%A6%E4%B8%8B%E9%99%8D%E6%B3%95%E6%98%AF%E4%BB%80%E4%B9%88
2.3 Momentum 是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#23-momentum-%E6%98%AF%E4%BB%80%E4%B9%88
2.4 SGD with Nesterov Acceleration 是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#24-sgd-with-nesterov-acceleration-%E6%98%AF%E4%BB%80%E4%B9%88
2.5 Adagrad 是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#25-adagrad-%E6%98%AF%E4%BB%80%E4%B9%88
2.6 RMSProp/AdaDelta 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#26-rmspropadadelta-%E6%98%AF%E4%BB%80%E4%B9%88
2.7 Adam 是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#27-adam-%E6%98%AF%E4%BB%80%E4%B9%88
2.8 Nadam 是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#28-nadam-%E6%98%AF%E4%BB%80%E4%B9%88
三、优化函数学霸笔记篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95%E5%8F%8A%E5%87%BD%E6%95%B0.md#%E4%B8%89%E4%BC%98%E5%8C%96%E5%87%BD%E6%95%B0%E5%AD%A6%E9%9C%B8%E7%AC%94%E8%AE%B0%E7%AF%87
【关于 归一化】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%BD%92%E4%B8%80%E5%8C%96.md
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%BD%92%E4%B8%80%E5%8C%96.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 为什么要归一化？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%BD%92%E4%B8%80%E5%8C%96.md#11-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E5%BD%92%E4%B8%80%E5%8C%96
二、介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%BD%92%E4%B8%80%E5%8C%96.md#%E4%BA%8C%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1 归一化 有 哪些方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%BD%92%E4%B8%80%E5%8C%96.md#21--%E5%BD%92%E4%B8%80%E5%8C%96-%E6%9C%89-%E5%93%AA%E4%BA%9B%E6%96%B9%E6%B3%95
2.2 归一化 各方法 特点？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%BD%92%E4%B8%80%E5%8C%96.md#22--%E5%BD%92%E4%B8%80%E5%8C%96-%E5%90%84%E6%96%B9%E6%B3%95-%E7%89%B9%E7%82%B9
2.3 归一化 的 意义？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%BD%92%E4%B8%80%E5%8C%96.md#23--%E5%BD%92%E4%B8%80%E5%8C%96-%E7%9A%84-%E6%84%8F%E4%B9%89
三、应用篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%BD%92%E4%B8%80%E5%8C%96.md#%E4%B8%89%E5%BA%94%E7%94%A8%E7%AF%87
3.1 哪些机器学习算法 需要做 归一化？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%BD%92%E4%B8%80%E5%8C%96.md#31-%E5%93%AA%E4%BA%9B%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E7%AE%97%E6%B3%95-%E9%9C%80%E8%A6%81%E5%81%9A-%E5%BD%92%E4%B8%80%E5%8C%96
3.2 哪些机器学习算法 不需要做 归一化？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%BD%92%E4%B8%80%E5%8C%96.md#32-%E5%93%AA%E4%BA%9B%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E7%AE%97%E6%B3%95-%E4%B8%8D%E9%9C%80%E8%A6%81%E5%81%9A-%E5%BD%92%E4%B8%80%E5%8C%96
【关于 判别式（discriminative）模型 vs. 生成式(generative)模型】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%88%A4%E5%88%AB%E5%BC%8Fvs%E7%94%9F%E6%88%90%E5%BC%8F.md
一、判别式模型篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%88%A4%E5%88%AB%E5%BC%8Fvs%E7%94%9F%E6%88%90%E5%BC%8F.md#%E4%B8%80%E5%88%A4%E5%88%AB%E5%BC%8F%E6%A8%A1%E5%9E%8B%E7%AF%87
1.1 什么是判别式模型？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%88%A4%E5%88%AB%E5%BC%8Fvs%E7%94%9F%E6%88%90%E5%BC%8F.md#11-%E4%BB%80%E4%B9%88%E6%98%AF%E5%88%A4%E5%88%AB%E5%BC%8F%E6%A8%A1%E5%9E%8B
1.2 判别式模型是思路是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%88%A4%E5%88%AB%E5%BC%8Fvs%E7%94%9F%E6%88%90%E5%BC%8F.md#12-%E5%88%A4%E5%88%AB%E5%BC%8F%E6%A8%A1%E5%9E%8B%E6%98%AF%E6%80%9D%E8%B7%AF%E6%98%AF%E4%BB%80%E4%B9%88
1.3 判别式模型的优点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%88%A4%E5%88%AB%E5%BC%8Fvs%E7%94%9F%E6%88%90%E5%BC%8F.md#13-%E5%88%A4%E5%88%AB%E5%BC%8F%E6%A8%A1%E5%9E%8B%E7%9A%84%E4%BC%98%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
二、生成式模型篇https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%88%A4%E5%88%AB%E5%BC%8Fvs%E7%94%9F%E6%88%90%E5%BC%8F.md#%E4%BA%8C%E7%94%9F%E6%88%90%E5%BC%8F%E6%A8%A1%E5%9E%8B%E7%AF%87
2.1 什么是生成式模型？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%88%A4%E5%88%AB%E5%BC%8Fvs%E7%94%9F%E6%88%90%E5%BC%8F.md#21-%E4%BB%80%E4%B9%88%E6%98%AF%E7%94%9F%E6%88%90%E5%BC%8F%E6%A8%A1%E5%9E%8B
2.2 生成式模型是思路是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%88%A4%E5%88%AB%E5%BC%8Fvs%E7%94%9F%E6%88%90%E5%BC%8F.md#22-%E7%94%9F%E6%88%90%E5%BC%8F%E6%A8%A1%E5%9E%8B%E6%98%AF%E6%80%9D%E8%B7%AF%E6%98%AF%E4%BB%80%E4%B9%88
2.3 生成式模型的优点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%88%A4%E5%88%AB%E5%BC%8Fvs%E7%94%9F%E6%88%90%E5%BC%8F.md#23-%E7%94%9F%E6%88%90%E5%BC%8F%E6%A8%A1%E5%9E%8B%E7%9A%84%E4%BC%98%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
2.4 生成式模型的缺点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/BasicAlgorithm/%E5%88%A4%E5%88%AB%E5%BC%8Fvs%E7%94%9F%E6%88%90%E5%BC%8F.md#24-%E7%94%9F%E6%88%90%E5%BC%8F%E6%A8%A1%E5%9E%8B%E7%9A%84%E7%BC%BA%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
【关于 机器学习算法篇】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm
https://github.com/turned123/NLP-Interview-Notes#二关于-机器学习算法篇那些你不知道的事
【关于 逻辑回归】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%80%BB%E8%BE%91%E5%9B%9E%E5%BD%92.md
一、介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%80%BB%E8%BE%91%E5%9B%9E%E5%BD%92.md#%E4%B8%80%E4%BB%8B%E7%BB%8D%E7%AF%87
1.1什么是逻辑回归https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%80%BB%E8%BE%91%E5%9B%9E%E5%BD%92.md#11%E4%BB%80%E4%B9%88%E6%98%AF%E9%80%BB%E8%BE%91%E5%9B%9E%E5%BD%92
1.2逻辑回归的优势https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%80%BB%E8%BE%91%E5%9B%9E%E5%BD%92.md#12%E9%80%BB%E8%BE%91%E5%9B%9E%E5%BD%92%E7%9A%84%E4%BC%98%E5%8A%BF
二、推导篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%80%BB%E8%BE%91%E5%9B%9E%E5%BD%92.md#%E4%BA%8C%E6%8E%A8%E5%AF%BC%E7%AF%87
2.1逻辑回归推导https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%80%BB%E8%BE%91%E5%9B%9E%E5%BD%92.md#21%E9%80%BB%E8%BE%91%E5%9B%9E%E5%BD%92%E6%8E%A8%E5%AF%BC
2.2求解优化https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%80%BB%E8%BE%91%E5%9B%9E%E5%BD%92.md#22%E6%B1%82%E8%A7%A3%E4%BC%98%E5%8C%96
【关于 支持向量机】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md
一、原理篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#%E4%B8%80%E5%8E%9F%E7%90%86%E7%AF%87
1.1 什么是SVM？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#11-%E4%BB%80%E4%B9%88%E6%98%AFsvm
Q.Ahttps://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#qa
1.2 SVM怎么发展的？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#12-svm%E6%80%8E%E4%B9%88%E5%8F%91%E5%B1%95%E7%9A%84
1.3 SVM存在什么问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#13-svm%E5%AD%98%E5%9C%A8%E4%BB%80%E4%B9%88%E9%97%AE%E9%A2%98
Q.Ahttps://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#qa-1
二、算法篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#%E4%BA%8C%E7%AE%97%E6%B3%95%E7%AF%87
2.1 什么是块算法？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#21-%E4%BB%80%E4%B9%88%E6%98%AF%E5%9D%97%E7%AE%97%E6%B3%95
2.2 什么是分解算法？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#22-%E4%BB%80%E4%B9%88%E6%98%AF%E5%88%86%E8%A7%A3%E7%AE%97%E6%B3%95
2.3 什么是序列最小优化算法？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#23-%E4%BB%80%E4%B9%88%E6%98%AF%E5%BA%8F%E5%88%97%E6%9C%80%E5%B0%8F%E4%BC%98%E5%8C%96%E7%AE%97%E6%B3%95
2.4 什么是增量算法？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#24-%E4%BB%80%E4%B9%88%E6%98%AF%E5%A2%9E%E9%87%8F%E7%AE%97%E6%B3%95
Q.Ahttps://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#qa-2
三、其他SVM篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#%E4%B8%89%E5%85%B6%E4%BB%96svm%E7%AF%87
3.1 什么是最小二次支持向量机？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#31-%E4%BB%80%E4%B9%88%E6%98%AF%E6%9C%80%E5%B0%8F%E4%BA%8C%E6%AC%A1%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA
3.2 什么是模糊支持向量机？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#32-%E4%BB%80%E4%B9%88%E6%98%AF%E6%A8%A1%E7%B3%8A%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA
3.3 什么是粒度支持向量机？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#33-%E4%BB%80%E4%B9%88%E6%98%AF%E7%B2%92%E5%BA%A6%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA
3.4 什么是多类训练算法？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#34-%E4%BB%80%E4%B9%88%E6%98%AF%E5%A4%9A%E7%B1%BB%E8%AE%AD%E7%BB%83%E7%AE%97%E6%B3%95
3.5 什么是孪生支持向量机？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#35-%E4%BB%80%E4%B9%88%E6%98%AF%E5%AD%AA%E7%94%9F%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA
3.6 什么是排序支持向量机？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#36-%E4%BB%80%E4%B9%88%E6%98%AF%E6%8E%92%E5%BA%8F%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA
Q.Ahttps://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#qa-3
四、应用篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#%E5%9B%9B%E5%BA%94%E7%94%A8%E7%AF%87
4.1 模式识别https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#41-%E6%A8%A1%E5%BC%8F%E8%AF%86%E5%88%AB
4.2 网页分类https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#42-%E7%BD%91%E9%A1%B5%E5%88%86%E7%B1%BB
4.3 系统建模与系统辨识https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#43-%E7%B3%BB%E7%BB%9F%E5%BB%BA%E6%A8%A1%E4%B8%8E%E7%B3%BB%E7%BB%9F%E8%BE%A8%E8%AF%86
4.4 其他https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#44-%E5%85%B6%E4%BB%96
五、对比篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#%E4%BA%94%E5%AF%B9%E6%AF%94%E7%AF%87
六、拓展篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E6%94%AF%E6%8C%81%E5%90%91%E9%87%8F%E6%9C%BA.md#%E5%85%AD%E6%8B%93%E5%B1%95%E7%AF%87
【关于 集成学习】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md
一、动机https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#%E4%B8%80%E5%8A%A8%E6%9C%BA
二、集成学习介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#%E4%BA%8C%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#21-%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1.1 集成学习的基本思想是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#211-%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0%E7%9A%84%E5%9F%BA%E6%9C%AC%E6%80%9D%E6%83%B3%E6%98%AF%E4%BB%80%E4%B9%88
2.1.2 集成学习为什么有效？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#212-%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0%E4%B8%BA%E4%BB%80%E4%B9%88%E6%9C%89%E6%95%88
三、 Boosting 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#%E4%B8%89-boosting-%E7%AF%87
3.1 用一句话概括 Boosting？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#31-%E7%94%A8%E4%B8%80%E5%8F%A5%E8%AF%9D%E6%A6%82%E6%8B%AC-boosting
3.2 Boosting 的特点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#32-boosting-%E7%9A%84%E7%89%B9%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
3.3 Boosting 的基本思想是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#33-boosting-%E7%9A%84%E5%9F%BA%E6%9C%AC%E6%80%9D%E6%83%B3%E6%98%AF%E4%BB%80%E4%B9%88
3.4 Boosting 的特点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#34-boosting-%E7%9A%84%E7%89%B9%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
3.5 GBDT 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#35-gbdt-%E6%98%AF%E4%BB%80%E4%B9%88
3.6 Xgboost 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#36-xgboost-%E6%98%AF%E4%BB%80%E4%B9%88
四、Bagging 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#%E5%9B%9Bbagging-%E7%AF%87
4.1 用一句话概括 Bagging？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#41-%E7%94%A8%E4%B8%80%E5%8F%A5%E8%AF%9D%E6%A6%82%E6%8B%AC-bagging
4.2 Bagging 的特点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#42-bagging-%E7%9A%84%E7%89%B9%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
4.3 Bagging 的基本思想是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#43-bagging-%E7%9A%84%E5%9F%BA%E6%9C%AC%E6%80%9D%E6%83%B3%E6%98%AF%E4%BB%80%E4%B9%88
4.4 Bagging 的基分类器如何选择？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#44-bagging-%E7%9A%84%E5%9F%BA%E5%88%86%E7%B1%BB%E5%99%A8%E5%A6%82%E4%BD%95%E9%80%89%E6%8B%A9
4.5 Bagging 的优点 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#45-bagging-%E7%9A%84%E4%BC%98%E7%82%B9-%E6%98%AF%E4%BB%80%E4%B9%88
4.6 Bagging 的特点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#46-bagging-%E7%9A%84%E7%89%B9%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
4.7 随机森林 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#47-%E9%9A%8F%E6%9C%BA%E6%A3%AE%E6%9E%97-%E6%98%AF%E4%BB%80%E4%B9%88
五、 Stacking 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#%E4%BA%94-stacking-%E7%AF%87
5.1 用一句话概括 Stacking ？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#51-%E7%94%A8%E4%B8%80%E5%8F%A5%E8%AF%9D%E6%A6%82%E6%8B%AC-stacking-
5.2 Stacking 的特点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#52-stacking-%E7%9A%84%E7%89%B9%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
5.3 Stacking 的基本思路是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#53-stacking-%E7%9A%84%E5%9F%BA%E6%9C%AC%E6%80%9D%E8%B7%AF%E6%98%AF%E4%BB%80%E4%B9%88
六、常见问题篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#%E5%85%AD%E5%B8%B8%E8%A7%81%E9%97%AE%E9%A2%98%E7%AF%87
6.1 为什么使用决策树作为基学习器？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#61-%E4%B8%BA%E4%BB%80%E4%B9%88%E4%BD%BF%E7%94%A8%E5%86%B3%E7%AD%96%E6%A0%91%E4%BD%9C%E4%B8%BA%E5%9F%BA%E5%AD%A6%E4%B9%A0%E5%99%A8
6.2 为什么不稳定的学习器更适合作为基学习器？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#62-%E4%B8%BA%E4%BB%80%E4%B9%88%E4%B8%8D%E7%A8%B3%E5%AE%9A%E7%9A%84%E5%AD%A6%E4%B9%A0%E5%99%A8%E6%9B%B4%E9%80%82%E5%90%88%E4%BD%9C%E4%B8%BA%E5%9F%BA%E5%AD%A6%E4%B9%A0%E5%99%A8
6.3 哪些模型适合作为基学习器？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#63-%E5%93%AA%E4%BA%9B%E6%A8%A1%E5%9E%8B%E9%80%82%E5%90%88%E4%BD%9C%E4%B8%BA%E5%9F%BA%E5%AD%A6%E4%B9%A0%E5%99%A8
6.4 Bagging 方法中能使用线性分类器作为基学习器吗？ Boosting 呢？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#64-bagging-%E6%96%B9%E6%B3%95%E4%B8%AD%E8%83%BD%E4%BD%BF%E7%94%A8%E7%BA%BF%E6%80%A7%E5%88%86%E7%B1%BB%E5%99%A8%E4%BD%9C%E4%B8%BA%E5%9F%BA%E5%AD%A6%E4%B9%A0%E5%99%A8%E5%90%97-boosting-%E5%91%A2
6.5 Boosting/Bagging 与 偏差/方差 的关系？https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#65-boostingbagging-%E4%B8%8E-%E5%81%8F%E5%B7%AE%E6%96%B9%E5%B7%AE-%E7%9A%84%E5%85%B3%E7%B3%BB
七、对比篇https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#%E4%B8%83%E5%AF%B9%E6%AF%94%E7%AF%87
7.1 LR vs GBDT?https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#71-lr-vs-gbdt
7.1.1 从机器学习三要素的角度https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#711-%E4%BB%8E%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E4%B8%89%E8%A6%81%E7%B4%A0%E7%9A%84%E8%A7%92%E5%BA%A6
7.1.1.1 从模型角度https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#7111-%E4%BB%8E%E6%A8%A1%E5%9E%8B%E8%A7%92%E5%BA%A6
7.1.1.2 从策略角度https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#7112-%E4%BB%8E%E7%AD%96%E7%95%A5%E8%A7%92%E5%BA%A6
7.1.1.2.1 从 Loss 角度https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#71121-%E4%BB%8E-loss-%E8%A7%92%E5%BA%A6
7.1.1.2.2 从 特征空间 角度https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#71122-%E4%BB%8E-%E7%89%B9%E5%BE%81%E7%A9%BA%E9%97%B4-%E8%A7%92%E5%BA%A6
7.1.1.2.3 从 正则 角度https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#71123-%E4%BB%8E-%E6%AD%A3%E5%88%99-%E8%A7%92%E5%BA%A6
7.1.1.3 从算法角度https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#7113-%E4%BB%8E%E7%AE%97%E6%B3%95%E8%A7%92%E5%BA%A6
7.1.2 从特征的角度https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#712-%E4%BB%8E%E7%89%B9%E5%BE%81%E7%9A%84%E8%A7%92%E5%BA%A6
7.1.2.1 特征组合https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#7121-%E7%89%B9%E5%BE%81%E7%BB%84%E5%90%88
7.1.2.2 特特征的稀疏性https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#7122-%E7%89%B9%E7%89%B9%E5%BE%81%E7%9A%84%E7%A8%80%E7%96%8F%E6%80%A7
7.1.3 数据假设不同https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#713--%E6%95%B0%E6%8D%AE%E5%81%87%E8%AE%BE%E4%B8%8D%E5%90%8C
7.1.3.1 LRhttps://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#7131-lr
7.1.3.2 GBDThttps://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#7132-gbdt
参考https://github.com/turned123/NLP-Interview-Notes/blob/main/MachineLearningAlgorithm/%E9%9B%86%E6%88%90%E5%AD%A6%E4%B9%A0.md#%E5%8F%82%E8%80%83
【关于 深度学习算法篇】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm
https://github.com/turned123/NLP-Interview-Notes#三关于-深度学习算法篇那些你不知道的事
【关于 CNN 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
二、CNN 卷积层篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#%E4%BA%8Ccnn-%E5%8D%B7%E7%A7%AF%E5%B1%82%E7%AF%87
2.1 卷积层的本质是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#21-%E5%8D%B7%E7%A7%AF%E5%B1%82%E7%9A%84%E6%9C%AC%E8%B4%A8%E6%98%AF%E4%BB%80%E4%B9%88
2.2 CNN 卷积层与全连接层的联系？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#22-cnn-%E5%8D%B7%E7%A7%AF%E5%B1%82%E4%B8%8E%E5%85%A8%E8%BF%9E%E6%8E%A5%E5%B1%82%E7%9A%84%E8%81%94%E7%B3%BB
2.3 channel的含义是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#23-channel%E7%9A%84%E5%90%AB%E4%B9%89%E6%98%AF%E4%BB%80%E4%B9%88
三、CNN 池化层篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#%E4%B8%89cnn-%E6%B1%A0%E5%8C%96%E5%B1%82%E7%AF%87
3.1 池化层针对区域是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#31-%E6%B1%A0%E5%8C%96%E5%B1%82%E9%92%88%E5%AF%B9%E5%8C%BA%E5%9F%9F%E6%98%AF%E4%BB%80%E4%B9%88
3.2 池化层的种类有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#32-%E6%B1%A0%E5%8C%96%E5%B1%82%E7%9A%84%E7%A7%8D%E7%B1%BB%E6%9C%89%E5%93%AA%E4%BA%9B
3.3 池化层的作用是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#33-%E6%B1%A0%E5%8C%96%E5%B1%82%E7%9A%84%E4%BD%9C%E7%94%A8%E6%98%AF%E4%BB%80%E4%B9%88
3.4 池化层 反向传播 是什么样的？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#34-%E6%B1%A0%E5%8C%96%E5%B1%82-%E5%8F%8D%E5%90%91%E4%BC%A0%E6%92%AD-%E6%98%AF%E4%BB%80%E4%B9%88%E6%A0%B7%E7%9A%84
3.5 mean pooling 池化层 反向传播 是什么样的？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#35-mean-pooling-%E6%B1%A0%E5%8C%96%E5%B1%82-%E5%8F%8D%E5%90%91%E4%BC%A0%E6%92%AD-%E6%98%AF%E4%BB%80%E4%B9%88%E6%A0%B7%E7%9A%84
3.6 max pooling 池化层 反向传播 是什么样的？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#36-max-pooling-%E6%B1%A0%E5%8C%96%E5%B1%82-%E5%8F%8D%E5%90%91%E4%BC%A0%E6%92%AD-%E6%98%AF%E4%BB%80%E4%B9%88%E6%A0%B7%E7%9A%84
四、CNN 整体篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#%E5%9B%9Bcnn-%E6%95%B4%E4%BD%93%E7%AF%87
4.1 CNN 的流程是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#41-cnn-%E7%9A%84%E6%B5%81%E7%A8%8B%E6%98%AF%E4%BB%80%E4%B9%88
4.2 CNN 的特点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#42-cnn-%E7%9A%84%E7%89%B9%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
4.3 卷积神经网络为什么会具有平移不变性？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#43-%E5%8D%B7%E7%A7%AF%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C%E4%B8%BA%E4%BB%80%E4%B9%88%E4%BC%9A%E5%85%B7%E6%9C%89%E5%B9%B3%E7%A7%BB%E4%B8%8D%E5%8F%98%E6%80%A7
4.4 卷积神经网络中im2col是如何实现的？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#44-%E5%8D%B7%E7%A7%AF%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C%E4%B8%ADim2col%E6%98%AF%E5%A6%82%E4%BD%95%E5%AE%9E%E7%8E%B0%E7%9A%84
4.5 CNN 的局限性是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#45-cnn-%E7%9A%84%E5%B1%80%E9%99%90%E6%80%A7%E6%98%AF%E4%BB%80%E4%B9%88
五、Iterated Dilated CNN 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#%E4%BA%94iterated-dilated-cnn-%E7%AF%87
5.1 什么是 Dilated CNN 空洞卷积？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#51-%E4%BB%80%E4%B9%88%E6%98%AF-dilated-cnn-%E7%A9%BA%E6%B4%9E%E5%8D%B7%E7%A7%AF
5.2 什么是 Iterated Dilated CNN？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#52-%E4%BB%80%E4%B9%88%E6%98%AF-iterated-dilated-cnn
六、反卷积 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#%E5%85%AD%E5%8F%8D%E5%8D%B7%E7%A7%AF-%E7%AF%87
6.1 解释反卷积的原理和用途？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/cnn/readme.md#61-%E8%A7%A3%E9%87%8A%E5%8F%8D%E5%8D%B7%E7%A7%AF%E7%9A%84%E5%8E%9F%E7%90%86%E5%92%8C%E7%94%A8%E9%80%94
【关于 Attention 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention
一、seq2seq 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#%E4%B8%80seq2seq-%E7%AF%87
1.1 seq2seq （Encoder-Decoder）是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#11-seq2seq-encoder-decoder%E6%98%AF%E4%BB%80%E4%B9%88
1.2 seq2seq 中 的 Encoder 怎么样？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#12-seq2seq-%E4%B8%AD-%E7%9A%84-encoder-%E6%80%8E%E4%B9%88%E6%A0%B7
1.3 seq2seq 中 的 Decoder 怎么样？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#13-seq2seq-%E4%B8%AD-%E7%9A%84-decoder-%E6%80%8E%E4%B9%88%E6%A0%B7
1.4 在 数学角度上 的 seq2seq ，你知道么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#14-%E5%9C%A8-%E6%95%B0%E5%AD%A6%E8%A7%92%E5%BA%A6%E4%B8%8A-%E7%9A%84-seq2seq-%E4%BD%A0%E7%9F%A5%E9%81%93%E4%B9%88
1.5 seq2seq 存在 什么 问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#15-seq2seq-%E5%AD%98%E5%9C%A8-%E4%BB%80%E4%B9%88-%E9%97%AE%E9%A2%98
二、Attention 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#%E4%BA%8Cattention-%E7%AF%87
2.1 什么是 Attention?https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#21-%E4%BB%80%E4%B9%88%E6%98%AF-attention
2.2 为什么引入 Attention机制？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#22-%E4%B8%BA%E4%BB%80%E4%B9%88%E5%BC%95%E5%85%A5-attention%E6%9C%BA%E5%88%B6
2.3 Attention 有什么作用？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#23-attention-%E6%9C%89%E4%BB%80%E4%B9%88%E4%BD%9C%E7%94%A8
2.4 Attention 流程是怎么样？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#24-attention-%E6%B5%81%E7%A8%8B%E6%98%AF%E6%80%8E%E4%B9%88%E6%A0%B7
步骤一 执行encoder (与 seq2seq 一致)https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#%E6%AD%A5%E9%AA%A4%E4%B8%80--%E6%89%A7%E8%A1%8Cencoder-%E4%B8%8E-seq2seq-%E4%B8%80%E8%87%B4
步骤二 计算对齐系数 ahttps://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#%E6%AD%A5%E9%AA%A4%E4%BA%8C--%E8%AE%A1%E7%AE%97%E5%AF%B9%E9%BD%90%E7%B3%BB%E6%95%B0-a
步骤三 计算上下文语义向量 Chttps://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#%E6%AD%A5%E9%AA%A4%E4%B8%89--%E8%AE%A1%E7%AE%97%E4%B8%8A%E4%B8%8B%E6%96%87%E8%AF%AD%E4%B9%89%E5%90%91%E9%87%8F-c
步骤四 更新decoder状态https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#%E6%AD%A5%E9%AA%A4%E5%9B%9B--%E6%9B%B4%E6%96%B0decoder%E7%8A%B6%E6%80%81
步骤五 计算输出预测词https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#%E6%AD%A5%E9%AA%A4%E4%BA%94-%E8%AE%A1%E7%AE%97%E8%BE%93%E5%87%BA%E9%A2%84%E6%B5%8B%E8%AF%8D
2.5 Attention 的应用领域有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#25-attention-%E7%9A%84%E5%BA%94%E7%94%A8%E9%A2%86%E5%9F%9F%E6%9C%89%E5%93%AA%E4%BA%9B
三、Attention 变体篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#%E4%B8%89attention-%E5%8F%98%E4%BD%93%E7%AF%87
3.1 Soft Attention 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#31-soft-attention-%E6%98%AF%E4%BB%80%E4%B9%88
3.2 Hard Attention 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#32-hard-attention-%E6%98%AF%E4%BB%80%E4%B9%88
3.3 Global Attention 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#33-global-attention-%E6%98%AF%E4%BB%80%E4%B9%88
3.4 Local Attention 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#34-local-attention-%E6%98%AF%E4%BB%80%E4%B9%88
3.5 self-attention 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/attention/readme.md#35-self-attention-%E6%98%AF%E4%BB%80%E4%B9%88
【关于 Transformer面试题】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer
【关于 Transformer】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 为什么要有 Transformer?https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#11-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E6%9C%89-transformer
1.2 Transformer 作用是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#12-transformer-%E4%BD%9C%E7%94%A8%E6%98%AF%E4%BB%80%E4%B9%88
二、整体结构篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#%E4%BA%8C%E6%95%B4%E4%BD%93%E7%BB%93%E6%9E%84%E7%AF%87
2.1 Transformer 整体结构是怎么样？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#21-transformer-%E6%95%B4%E4%BD%93%E7%BB%93%E6%9E%84%E6%98%AF%E6%80%8E%E4%B9%88%E6%A0%B7
2.2 Transformer-encoder 结构怎么样？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#22-transformer-encoder-%E7%BB%93%E6%9E%84%E6%80%8E%E4%B9%88%E6%A0%B7
2.3 Transformer-decoder 结构怎么样?https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#23-transformer-decoder-%E7%BB%93%E6%9E%84%E6%80%8E%E4%B9%88%E6%A0%B7
三、模块篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#%E4%B8%89%E6%A8%A1%E5%9D%97%E7%AF%87
3.1 self-attention 模块https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#31-self-attention-%E6%A8%A1%E5%9D%97
3.1.1 传统 attention 是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#311-%E4%BC%A0%E7%BB%9F-attention-%E6%98%AF%E4%BB%80%E4%B9%88
3.1.2 为什么 会有self-attention?https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#312-%E4%B8%BA%E4%BB%80%E4%B9%88-%E4%BC%9A%E6%9C%89self-attention
3.1.3 self-attention 的核心思想是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#313-self-attention-%E7%9A%84%E6%A0%B8%E5%BF%83%E6%80%9D%E6%83%B3%E6%98%AF%E4%BB%80%E4%B9%88
3.1.4 self-attention 的目的是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#314-self-attention-%E7%9A%84%E7%9B%AE%E7%9A%84%E6%98%AF%E4%BB%80%E4%B9%88
3.1.5 self-attention 的怎么计算的?https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#315-self-attention-%E7%9A%84%E6%80%8E%E4%B9%88%E8%AE%A1%E7%AE%97%E7%9A%84
3.1.6 self-attention 为什么Q和K使用不同的权重矩阵生成，为何不能使用同一个值进行自身的点乘？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#316-self-attention-%E4%B8%BA%E4%BB%80%E4%B9%88q%E5%92%8Ck%E4%BD%BF%E7%94%A8%E4%B8%8D%E5%90%8C%E7%9A%84%E6%9D%83%E9%87%8D%E7%9F%A9%E9%98%B5%E7%94%9F%E6%88%90%E4%B8%BA%E4%BD%95%E4%B8%8D%E8%83%BD%E4%BD%BF%E7%94%A8%E5%90%8C%E4%B8%80%E4%B8%AA%E5%80%BC%E8%BF%9B%E8%A1%8C%E8%87%AA%E8%BA%AB%E7%9A%84%E7%82%B9%E4%B9%98
3.1.7 为什么采用点积模型的 self-attention 而不采用加性模型？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#317-%E4%B8%BA%E4%BB%80%E4%B9%88%E9%87%87%E7%94%A8%E7%82%B9%E7%A7%AF%E6%A8%A1%E5%9E%8B%E7%9A%84-self-attention-%E8%80%8C%E4%B8%8D%E9%87%87%E7%94%A8%E5%8A%A0%E6%80%A7%E6%A8%A1%E5%9E%8B
3.1.8 Transformer 中在计算 self-attention 时为什么要除以 $\sqrt{d}$？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#318-transformer-%E4%B8%AD%E5%9C%A8%E8%AE%A1%E7%AE%97-self-attention-%E6%97%B6%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E9%99%A4%E4%BB%A5-sqrtd
3.1.9 self-attention 如何解决长距离依赖问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#319-self-attention-%E5%A6%82%E4%BD%95%E8%A7%A3%E5%86%B3%E9%95%BF%E8%B7%9D%E7%A6%BB%E4%BE%9D%E8%B5%96%E9%97%AE%E9%A2%98
3.1.10 self-attention 如何并行化？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#3110-self-attention-%E5%A6%82%E4%BD%95%E5%B9%B6%E8%A1%8C%E5%8C%96
3.2 multi-head attention 模块https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#32-multi-head-attention-%E6%A8%A1%E5%9D%97
3.2.1 multi-head attention 的思路是什么样?https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#321-multi-head-attention-%E7%9A%84%E6%80%9D%E8%B7%AF%E6%98%AF%E4%BB%80%E4%B9%88%E6%A0%B7
3.2.2 multi-head attention 的步骤是什么样?https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#322-multi-head-attention-%E7%9A%84%E6%AD%A5%E9%AA%A4%E6%98%AF%E4%BB%80%E4%B9%88%E6%A0%B7
3.2.3 Transformer为何使用多头注意力机制？（为什么不使用一个头）https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#323-transformer%E4%B8%BA%E4%BD%95%E4%BD%BF%E7%94%A8%E5%A4%9A%E5%A4%B4%E6%B3%A8%E6%84%8F%E5%8A%9B%E6%9C%BA%E5%88%B6%E4%B8%BA%E4%BB%80%E4%B9%88%E4%B8%8D%E4%BD%BF%E7%94%A8%E4%B8%80%E4%B8%AA%E5%A4%B4
3.2.4 为什么在进行多头注意力的时候需要对每个head进行降维？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#324-%E4%B8%BA%E4%BB%80%E4%B9%88%E5%9C%A8%E8%BF%9B%E8%A1%8C%E5%A4%9A%E5%A4%B4%E6%B3%A8%E6%84%8F%E5%8A%9B%E7%9A%84%E6%97%B6%E5%80%99%E9%9C%80%E8%A6%81%E5%AF%B9%E6%AF%8F%E4%B8%AAhead%E8%BF%9B%E8%A1%8C%E9%99%8D%E7%BB%B4
3.2.5 multi-head attention 代码介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#325-multi-head-attention-%E4%BB%A3%E7%A0%81%E4%BB%8B%E7%BB%8D
3.3 位置编码（Position encoding）模块https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#33-%E4%BD%8D%E7%BD%AE%E7%BC%96%E7%A0%81position-encoding%E6%A8%A1%E5%9D%97
3.3.1 为什么要 加入 位置编码（Position encoding） ？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#331-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81-%E5%8A%A0%E5%85%A5-%E4%BD%8D%E7%BD%AE%E7%BC%96%E7%A0%81position-encoding-
3.3.2 位置编码（Position encoding）的思路是什么 ？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#332-%E4%BD%8D%E7%BD%AE%E7%BC%96%E7%A0%81position-encoding%E7%9A%84%E6%80%9D%E8%B7%AF%E6%98%AF%E4%BB%80%E4%B9%88-
3.3.3 位置编码（Position encoding）的作用是什么 ？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#333-%E4%BD%8D%E7%BD%AE%E7%BC%96%E7%A0%81position-encoding%E7%9A%84%E4%BD%9C%E7%94%A8%E6%98%AF%E4%BB%80%E4%B9%88-
3.3.4 位置编码（Position encoding）的步骤是什么 ？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#334-%E4%BD%8D%E7%BD%AE%E7%BC%96%E7%A0%81position-encoding%E7%9A%84%E6%AD%A5%E9%AA%A4%E6%98%AF%E4%BB%80%E4%B9%88-
3.3.5 Position encoding为什么选择相加而不是拼接呢？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#335-position-encoding%E4%B8%BA%E4%BB%80%E4%B9%88%E9%80%89%E6%8B%A9%E7%9B%B8%E5%8A%A0%E8%80%8C%E4%B8%8D%E6%98%AF%E6%8B%BC%E6%8E%A5%E5%91%A2
3.3.6 Position encoding和 Position embedding的区别？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#336-position-encoding%E5%92%8C-position-embedding%E7%9A%84%E5%8C%BA%E5%88%AB
3.3.7 为何17年提出Transformer时采用的是 Position Encoder 而不是Position Embedding？而Bert却采用的是 Position Embedding ？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#337-%E4%B8%BA%E4%BD%9517%E5%B9%B4%E6%8F%90%E5%87%BAtransformer%E6%97%B6%E9%87%87%E7%94%A8%E7%9A%84%E6%98%AF-position-encoder--%E8%80%8C%E4%B8%8D%E6%98%AFposition-embedding%E8%80%8Cbert%E5%8D%B4%E9%87%87%E7%94%A8%E7%9A%84%E6%98%AF-position-embedding-
3.3.8 位置编码（Position encoding）的代码介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#338-%E4%BD%8D%E7%BD%AE%E7%BC%96%E7%A0%81position-encoding%E7%9A%84%E4%BB%A3%E7%A0%81%E4%BB%8B%E7%BB%8D
3.4 残差模块模块https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#34-%E6%AE%8B%E5%B7%AE%E6%A8%A1%E5%9D%97%E6%A8%A1%E5%9D%97
3.4.1 为什么要 加入 残差模块？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#341-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81-%E5%8A%A0%E5%85%A5-%E6%AE%8B%E5%B7%AE%E6%A8%A1%E5%9D%97
3.5 Layer normalization 模块https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#35-layer-normalization-%E6%A8%A1%E5%9D%97
3.5.1 为什么要 加入 Layer normalization 模块？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#351-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81-%E5%8A%A0%E5%85%A5-layer-normalization-%E6%A8%A1%E5%9D%97
3.5.2 Layer normalization 模块的是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#352-layer-normalization-%E6%A8%A1%E5%9D%97%E7%9A%84%E6%98%AF%E4%BB%80%E4%B9%88
3.5.3 Batch normalization 和 Layer normalization 的区别？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#353-batch-normalization-%E5%92%8C-layer-normalization-%E7%9A%84%E5%8C%BA%E5%88%AB
3.5.4 Transformer 中为什么要舍弃 Batch normalization 改用 Layer normalization 呢?https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#354-transformer-%E4%B8%AD%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E8%88%8D%E5%BC%83-batch-normalization-%E6%94%B9%E7%94%A8-layer-normalization-%E5%91%A2
3.5.5 Layer normalization 模块代码介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#355--layer-normalization-%E6%A8%A1%E5%9D%97%E4%BB%A3%E7%A0%81%E4%BB%8B%E7%BB%8D
3.6 Mask 模块https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#36-mask-%E6%A8%A1%E5%9D%97
3.6.1 什么是 Mask？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#361-%E4%BB%80%E4%B9%88%E6%98%AF-mask
3.6.2 Transformer 中用到 几种 Mask？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#362-transformer-%E4%B8%AD%E7%94%A8%E5%88%B0-%E5%87%A0%E7%A7%8D-mask
3.6.3 能不能介绍一下 Transformer 中用到几种 Mask？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/readme.md#363-%E8%83%BD%E4%B8%8D%E8%83%BD%E4%BB%8B%E7%BB%8D%E4%B8%80%E4%B8%8B-transformer-%E4%B8%AD%E7%94%A8%E5%88%B0%E5%87%A0%E7%A7%8D-mask
【关于 Transformer 问题及改进】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md
一、Transformer 问题篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md#%E4%B8%80transformer-%E9%97%AE%E9%A2%98%E7%AF%87
1.1 既然 Transformer 怎么牛逼，是否还存在一些问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md#11-%E6%97%A2%E7%84%B6-transformer-%E6%80%8E%E4%B9%88%E7%89%9B%E9%80%BC%E6%98%AF%E5%90%A6%E8%BF%98%E5%AD%98%E5%9C%A8%E4%B8%80%E4%BA%9B%E9%97%AE%E9%A2%98
二、每个问题的解决方法是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md#%E4%BA%8C%E6%AF%8F%E4%B8%AA%E9%97%AE%E9%A2%98%E7%9A%84%E8%A7%A3%E5%86%B3%E6%96%B9%E6%B3%95%E6%98%AF%E4%BB%80%E4%B9%88
2.1 问题一：Transformer 不能很好的处理超长输入问题https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md#21-%E9%97%AE%E9%A2%98%E4%B8%80transformer-%E4%B8%8D%E8%83%BD%E5%BE%88%E5%A5%BD%E7%9A%84%E5%A4%84%E7%90%86%E8%B6%85%E9%95%BF%E8%BE%93%E5%85%A5%E9%97%AE%E9%A2%98
2.1.1 Transformer 固定了句子长度？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md#211-transformer-%E5%9B%BA%E5%AE%9A%E4%BA%86%E5%8F%A5%E5%AD%90%E9%95%BF%E5%BA%A6
2.1.2 Transformer 固定了句子长度 的目的是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md#212-transformer-%E5%9B%BA%E5%AE%9A%E4%BA%86%E5%8F%A5%E5%AD%90%E9%95%BF%E5%BA%A6-%E7%9A%84%E7%9B%AE%E7%9A%84%E6%98%AF%E4%BB%80%E4%B9%88
2.1.3 Transformer 针对该问题的处理方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md#213-transformer-%E9%92%88%E5%AF%B9%E8%AF%A5%E9%97%AE%E9%A2%98%E7%9A%84%E5%A4%84%E7%90%86%E6%96%B9%E6%B3%95
2.2 问题二：Transformer 方向信息以及相对位置 的 缺失 问题https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md#22-%E9%97%AE%E9%A2%98%E4%BA%8Ctransformer-%E6%96%B9%E5%90%91%E4%BF%A1%E6%81%AF%E4%BB%A5%E5%8F%8A%E7%9B%B8%E5%AF%B9%E4%BD%8D%E7%BD%AE-%E7%9A%84-%E7%BC%BA%E5%A4%B1-%E9%97%AE%E9%A2%98
2.3 问题三：缺少Recurrent Inductive Biashttps://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md#23--%E9%97%AE%E9%A2%98%E4%B8%89%E7%BC%BA%E5%B0%91recurrent-inductive-bias
问题四：问题四：Transformer是非图灵完备的： 非图灵完备通俗的理解，就是无法解决所有的问题https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md#%E9%97%AE%E9%A2%98%E5%9B%9B%E9%97%AE%E9%A2%98%E5%9B%9Btransformer%E6%98%AF%E9%9D%9E%E5%9B%BE%E7%81%B5%E5%AE%8C%E5%A4%87%E7%9A%84-%E9%9D%9E%E5%9B%BE%E7%81%B5%E5%AE%8C%E5%A4%87%E9%80%9A%E4%BF%97%E7%9A%84%E7%90%86%E8%A7%A3%E5%B0%B1%E6%98%AF%E6%97%A0%E6%B3%95%E8%A7%A3%E5%86%B3%E6%89%80%E6%9C%89%E7%9A%84%E9%97%AE%E9%A2%98
问题五：transformer缺少conditional computation；https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md#%E9%97%AE%E9%A2%98%E4%BA%94transformer%E7%BC%BA%E5%B0%91conditional-computation
问题六：transformer 时间复杂度 和 空间复杂度 过大问题；https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/transformer/transformer_error.md#%E9%97%AE%E9%A2%98%E5%85%ADtransformer-%E6%97%B6%E9%97%B4%E5%A4%8D%E6%9D%82%E5%BA%A6-%E5%92%8C-%E7%A9%BA%E9%97%B4%E5%A4%8D%E6%9D%82%E5%BA%A6-%E8%BF%87%E5%A4%A7%E9%97%AE%E9%A2%98
【关于 生成对抗网络 GAN 】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md
一、动机https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA
二、介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md#%E4%BA%8C%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1 GAN 的基本思想https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md#21-gan-%E7%9A%84%E5%9F%BA%E6%9C%AC%E6%80%9D%E6%83%B3
2.2 GAN 基本介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md#22-gan-%E5%9F%BA%E6%9C%AC%E4%BB%8B%E7%BB%8D
2.2.1 GAN 的基本结构https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md#221--gan-%E7%9A%84%E5%9F%BA%E6%9C%AC%E7%BB%93%E6%9E%84
2.2.2 GAN 的基本思想https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md#222-gan-%E7%9A%84%E5%9F%BA%E6%9C%AC%E6%80%9D%E6%83%B3
三、训练篇https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md#%E4%B8%89%E8%AE%AD%E7%BB%83%E7%AF%87
3.1 生成器介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md#31-%E7%94%9F%E6%88%90%E5%99%A8%E4%BB%8B%E7%BB%8D
3.2 判别器介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md#32-%E5%88%A4%E5%88%AB%E5%99%A8%E4%BB%8B%E7%BB%8D
3.3 训练过程https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md#33-%E8%AE%AD%E7%BB%83%E8%BF%87%E7%A8%8B
3.4 训练所涉及相关理论基础https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md#34--%E8%AE%AD%E7%BB%83%E6%89%80%E6%B6%89%E5%8F%8A%E7%9B%B8%E5%85%B3%E7%90%86%E8%AE%BA%E5%9F%BA%E7%A1%80
四、总结https://github.com/turned123/NLP-Interview-Notes/blob/main/DeepLearningAlgorithm/adversarial_training_study/readme.md#%E5%9B%9B%E6%80%BB%E7%BB%93
【关于 NLP 学习算法】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview
https://github.com/turned123/NLP-Interview-Notes#四关于-nlp-学习算法那些你不知道的事
【关于 信息抽取】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/NER
https://github.com/turned123/NLP-Interview-Notes#41-关于-信息抽取那些你不知道的事
【关于 命名实体识别】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/NER
https://github.com/turned123/NLP-Interview-Notes#411-关于-命名实体识别那些你不知道的事
【关于 HMM->MEMM->CRF】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf
一、基础信息 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#%E4%B8%80%E5%9F%BA%E7%A1%80%E4%BF%A1%E6%81%AF-%E4%BB%8B%E7%BB%8D%E7%AF%87
1.1 什么是概率图模型？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#11-%E4%BB%80%E4%B9%88%E6%98%AF%E6%A6%82%E7%8E%87%E5%9B%BE%E6%A8%A1%E5%9E%8B
1.2 什么是 随机场？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#12-%E4%BB%80%E4%B9%88%E6%98%AF-%E9%9A%8F%E6%9C%BA%E5%9C%BA
二、马尔可夫过程 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#%E4%BA%8C%E9%A9%AC%E5%B0%94%E5%8F%AF%E5%A4%AB%E8%BF%87%E7%A8%8B-%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1 什么是 马尔可夫过程？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#21-%E4%BB%80%E4%B9%88%E6%98%AF-%E9%A9%AC%E5%B0%94%E5%8F%AF%E5%A4%AB%E8%BF%87%E7%A8%8B
2.2 马尔可夫过程 的核心思想 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#22-%E9%A9%AC%E5%B0%94%E5%8F%AF%E5%A4%AB%E8%BF%87%E7%A8%8B-%E7%9A%84%E6%A0%B8%E5%BF%83%E6%80%9D%E6%83%B3-%E6%98%AF%E4%BB%80%E4%B9%88
三、隐马尔科夫算法 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#%E4%B8%89%E9%9A%90%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E7%AE%97%E6%B3%95-%E7%AF%87
3.1 隐马尔科夫算法 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#31-%E9%9A%90%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E7%AE%97%E6%B3%95-%E4%BB%8B%E7%BB%8D%E7%AF%87
3.1.1 隐马尔科夫算法 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#311-%E9%9A%90%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E7%AE%97%E6%B3%95-%E6%98%AF%E4%BB%80%E4%B9%88
3.1.2 隐马尔科夫算法 中 两个序列 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#312-%E9%9A%90%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E7%AE%97%E6%B3%95-%E4%B8%AD-%E4%B8%A4%E4%B8%AA%E5%BA%8F%E5%88%97-%E6%98%AF%E4%BB%80%E4%B9%88
3.1.3 隐马尔科夫算法 中 三个矩阵 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#313-%E9%9A%90%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E7%AE%97%E6%B3%95-%E4%B8%AD-%E4%B8%89%E4%B8%AA%E7%9F%A9%E9%98%B5-%E6%98%AF%E4%BB%80%E4%B9%88
3.1.4 隐马尔科夫算法 中 两个假设 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#314-%E9%9A%90%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E7%AE%97%E6%B3%95-%E4%B8%AD-%E4%B8%A4%E4%B8%AA%E5%81%87%E8%AE%BE-%E6%98%AF%E4%BB%80%E4%B9%88
3.1.5 隐马尔科夫算法 中 工作流程 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#315-%E9%9A%90%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E7%AE%97%E6%B3%95-%E4%B8%AD-%E5%B7%A5%E4%BD%9C%E6%B5%81%E7%A8%8B-%E6%98%AF%E4%BB%80%E4%B9%88
3.2 隐马尔科夫算法 模型计算过程篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#32-%E9%9A%90%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E7%AE%97%E6%B3%95-%E6%A8%A1%E5%9E%8B%E8%AE%A1%E7%AE%97%E8%BF%87%E7%A8%8B%E7%AF%87
3.2.1 隐马尔科夫算法 学习训练过程 是什么样的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#321-%E9%9A%90%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E7%AE%97%E6%B3%95-%E5%AD%A6%E4%B9%A0%E8%AE%AD%E7%BB%83%E8%BF%87%E7%A8%8B-%E6%98%AF%E4%BB%80%E4%B9%88%E6%A0%B7%E7%9A%84
3.2.2 隐马尔科夫算法 序列标注（解码）过程 是什么样的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#322-%E9%9A%90%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E7%AE%97%E6%B3%95-%E5%BA%8F%E5%88%97%E6%A0%87%E6%B3%A8%E8%A7%A3%E7%A0%81%E8%BF%87%E7%A8%8B-%E6%98%AF%E4%BB%80%E4%B9%88%E6%A0%B7%E7%9A%84
3.2.3 隐马尔科夫算法 序列概率过程 是什么样的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#323-%E9%9A%90%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E7%AE%97%E6%B3%95-%E5%BA%8F%E5%88%97%E6%A6%82%E7%8E%87%E8%BF%87%E7%A8%8B-%E6%98%AF%E4%BB%80%E4%B9%88%E6%A0%B7%E7%9A%84
3.3 隐马尔科夫算法 问题篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#33-%E9%9A%90%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E7%AE%97%E6%B3%95-%E9%97%AE%E9%A2%98%E7%AF%87
四、最大熵马尔科夫模型（MEMM）篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#%E5%9B%9B%E6%9C%80%E5%A4%A7%E7%86%B5%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E6%A8%A1%E5%9E%8Bmemm%E7%AF%87
4.1 最大熵马尔科夫模型（MEMM）动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#41-%E6%9C%80%E5%A4%A7%E7%86%B5%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E6%A8%A1%E5%9E%8Bmemm%E5%8A%A8%E6%9C%BA%E7%AF%87
4.1.1 HMM 存在 什么问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#411-hmm-%E5%AD%98%E5%9C%A8-%E4%BB%80%E4%B9%88%E9%97%AE%E9%A2%98
4.2 最大熵马尔科夫模型（MEMM）介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#42-%E6%9C%80%E5%A4%A7%E7%86%B5%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E6%A8%A1%E5%9E%8Bmemm%E4%BB%8B%E7%BB%8D%E7%AF%87
4.2.1 最大熵马尔科夫模型（MEMM） 是什么样？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#421-%E6%9C%80%E5%A4%A7%E7%86%B5%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E6%A8%A1%E5%9E%8Bmemm-%E6%98%AF%E4%BB%80%E4%B9%88%E6%A0%B7
4.2.2 最大熵马尔科夫模型（MEMM） 如何解决 HMM 问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#422-%E6%9C%80%E5%A4%A7%E7%86%B5%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E6%A8%A1%E5%9E%8Bmemm-%E5%A6%82%E4%BD%95%E8%A7%A3%E5%86%B3-hmm-%E9%97%AE%E9%A2%98
4.3 最大熵马尔科夫模型（MEMM）问题篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#43-%E6%9C%80%E5%A4%A7%E7%86%B5%E9%A9%AC%E5%B0%94%E7%A7%91%E5%A4%AB%E6%A8%A1%E5%9E%8Bmemm%E9%97%AE%E9%A2%98%E7%AF%87
五、条件随机场（CRF）篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#%E4%BA%94%E6%9D%A1%E4%BB%B6%E9%9A%8F%E6%9C%BA%E5%9C%BAcrf%E7%AF%87
5.1 CRF 动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#51-crf-%E5%8A%A8%E6%9C%BA%E7%AF%87
5.1.1 HMM 和 MEMM 存在什么问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#511-hmm-%E5%92%8C-memm-%E5%AD%98%E5%9C%A8%E4%BB%80%E4%B9%88%E9%97%AE%E9%A2%98
5.2 CRF 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#52-crf-%E4%BB%8B%E7%BB%8D%E7%AF%87
5.2.1 什么是 CRF?https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#521-%E4%BB%80%E4%B9%88%E6%98%AF-crf
5.2.2 CRF 的 主要思想是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#522-crf-%E7%9A%84-%E4%B8%BB%E8%A6%81%E6%80%9D%E6%83%B3%E6%98%AF%E4%BB%80%E4%B9%88
5.2.3 CRF 的定义是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#523--crf-%E7%9A%84%E5%AE%9A%E4%B9%89%E6%98%AF%E4%BB%80%E4%B9%88
5.2.4 CRF 的 流程是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#524-crf-%E7%9A%84-%E6%B5%81%E7%A8%8B%E6%98%AF%E4%BB%80%E4%B9%88
5.3 CRF 优缺点篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#53-crf-%E4%BC%98%E7%BC%BA%E7%82%B9%E7%AF%87
5.3.1 CRF 的 优点在哪里？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#531-crf-%E7%9A%84-%E4%BC%98%E7%82%B9%E5%9C%A8%E5%93%AA%E9%87%8C
5.3.2 CRF 的 缺点在哪里？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#532-crf-%E7%9A%84-%E7%BC%BA%E7%82%B9%E5%9C%A8%E5%93%AA%E9%87%8C
5.4 CRF 复现？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#54-crf-%E5%A4%8D%E7%8E%B0
六、对比篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#%E5%85%AD%E5%AF%B9%E6%AF%94%E7%AF%87
6.1 CRF模型 和 HMM和MEMM模型 区别？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/crf/readme.md#61-crf%E6%A8%A1%E5%9E%8B-%E5%92%8C-hmm%E5%92%8Cmemm%E6%A8%A1%E5%9E%8B-%E5%8C%BA%E5%88%AB
【关于 DNN-CRF】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md
一、基本信息https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#%E4%B8%80%E5%9F%BA%E6%9C%AC%E4%BF%A1%E6%81%AF
1.1 命名实体识别 评价指标 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#11-%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB-%E8%AF%84%E4%BB%B7%E6%8C%87%E6%A0%87-%E6%98%AF%E4%BB%80%E4%B9%88
二、传统的命名实体识别方法https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#%E4%BA%8C%E4%BC%A0%E7%BB%9F%E7%9A%84%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E6%96%B9%E6%B3%95
2.1 基于规则的命名实体识别方法是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#21-%E5%9F%BA%E4%BA%8E%E8%A7%84%E5%88%99%E7%9A%84%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E6%96%B9%E6%B3%95%E6%98%AF%E4%BB%80%E4%B9%88
2.2 基于无监督学习的命名实体识别方法是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#22-%E5%9F%BA%E4%BA%8E%E6%97%A0%E7%9B%91%E7%9D%A3%E5%AD%A6%E4%B9%A0%E7%9A%84%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E6%96%B9%E6%B3%95%E6%98%AF%E4%BB%80%E4%B9%88
2.3 基于特征的监督学习的命名实体识别方法是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#23-%E5%9F%BA%E4%BA%8E%E7%89%B9%E5%BE%81%E7%9A%84%E7%9B%91%E7%9D%A3%E5%AD%A6%E4%B9%A0%E7%9A%84%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E6%96%B9%E6%B3%95%E6%98%AF%E4%BB%80%E4%B9%88
三、基于深度学习的命名实体识别方法https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#%E4%B8%89%E5%9F%BA%E4%BA%8E%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%9A%84%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E6%96%B9%E6%B3%95
3.1 基于深度学习的命名实体识别方法 相比于 基于机器学习的命名实体识别方法的优点？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#31-%E5%9F%BA%E4%BA%8E%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%9A%84%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E6%96%B9%E6%B3%95-%E7%9B%B8%E6%AF%94%E4%BA%8E-%E5%9F%BA%E4%BA%8E%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E7%9A%84%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E6%96%B9%E6%B3%95%E7%9A%84%E4%BC%98%E7%82%B9
3.2 基于深度学习的命名实体识别方法 的 结构是怎么样？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#32-%E5%9F%BA%E4%BA%8E%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%9A%84%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E6%96%B9%E6%B3%95--%E7%9A%84-%E7%BB%93%E6%9E%84%E6%98%AF%E6%80%8E%E4%B9%88%E6%A0%B7
3.3 分布式输入层 是什么，有哪些方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#33-%E5%88%86%E5%B8%83%E5%BC%8F%E8%BE%93%E5%85%A5%E5%B1%82-%E6%98%AF%E4%BB%80%E4%B9%88%E6%9C%89%E5%93%AA%E4%BA%9B%E6%96%B9%E6%B3%95
3.4 文本编码器篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#34-%E6%96%87%E6%9C%AC%E7%BC%96%E7%A0%81%E5%99%A8%E7%AF%87
3.4.1 BiLSTM-CRF 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#341-bilstm-crf-%E7%AF%87
3.4.1.1 什么是 BiLSTM-CRF？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#3411-%E4%BB%80%E4%B9%88%E6%98%AF-bilstm-crf
3.4.1.2 为什么要用 BiLSTM？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#3412-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E7%94%A8-bilstm
3.4.2 IDCNN-CRF 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#342-idcnn-crf-%E7%AF%87
3.4.2.1 什么是 Dilated CNN？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#3421-%E4%BB%80%E4%B9%88%E6%98%AF-dilated-cnn
3.4.2.2 为什么会有 Dilated CNN？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#3422-%E4%B8%BA%E4%BB%80%E4%B9%88%E4%BC%9A%E6%9C%89-dilated-cnn
3.4.2.3 Dilated CNN 的优点？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#3423-dilated-cnn-%E7%9A%84%E4%BC%98%E7%82%B9
3.4.2.4 IDCNN-CRF 介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#3424-idcnn-crf-%E4%BB%8B%E7%BB%8D
3.5 标签解码器篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#35-%E6%A0%87%E7%AD%BE%E8%A7%A3%E7%A0%81%E5%99%A8%E7%AF%87
3.5.1 标签解码器是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#351-%E6%A0%87%E7%AD%BE%E8%A7%A3%E7%A0%81%E5%99%A8%E6%98%AF%E4%BB%80%E4%B9%88
3.5.2 MLP+softmax层 介绍？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#352-mlpsoftmax%E5%B1%82-%E4%BB%8B%E7%BB%8D
3.5.3 条件随机场CRF层 介绍？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#353-%E6%9D%A1%E4%BB%B6%E9%9A%8F%E6%9C%BA%E5%9C%BAcrf%E5%B1%82-%E4%BB%8B%E7%BB%8D
3.5.4 循环神经网络RNN层 介绍？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#354-%E5%BE%AA%E7%8E%AF%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9Crnn%E5%B1%82-%E4%BB%8B%E7%BB%8D
3.5.3 指针网路层 介绍？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#353-%E6%8C%87%E9%92%88%E7%BD%91%E8%B7%AF%E5%B1%82-%E4%BB%8B%E7%BB%8D
四、对比 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#%E5%9B%9B%E5%AF%B9%E6%AF%94-%E7%AF%87
4.1 CNN-CRF vs BiLSTM-CRF vs IDCNN-CRF?https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#41-cnn-crf-vs-bilstm-crf-vs-idcnn-crf
4.2 为什么 DNN 后面要加 CRF?https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#42-%E4%B8%BA%E4%BB%80%E4%B9%88-dnn-%E5%90%8E%E9%9D%A2%E8%A6%81%E5%8A%A0-crf
4.3 CRF in TensorFlow V.S. CRF in discrete toolkit？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/DNN/readme.md#43-crf-in-tensorflow-vs-crf-in-discrete-toolkit
【关于 中文领域 NER】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 中文命名实体识别 与 英文命名实体识别的区别？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#11-%E4%B8%AD%E6%96%87%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB-%E4%B8%8E-%E8%8B%B1%E6%96%87%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E7%9A%84%E5%8C%BA%E5%88%AB
二、词汇增强篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#%E4%BA%8C%E8%AF%8D%E6%B1%87%E5%A2%9E%E5%BC%BA%E7%AF%87
2.1 什么是 词汇增强？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#21-%E4%BB%80%E4%B9%88%E6%98%AF-%E8%AF%8D%E6%B1%87%E5%A2%9E%E5%BC%BA
2.2 为什么说 「词汇增强」 方法对于中文 NER 任务有效呢？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#22-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%AF%B4-%E8%AF%8D%E6%B1%87%E5%A2%9E%E5%BC%BA-%E6%96%B9%E6%B3%95%E5%AF%B9%E4%BA%8E%E4%B8%AD%E6%96%87-ner-%E4%BB%BB%E5%8A%A1%E6%9C%89%E6%95%88%E5%91%A2
2.3 词汇增强 方法有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#23-%E8%AF%8D%E6%B1%87%E5%A2%9E%E5%BC%BA-%E6%96%B9%E6%B3%95%E6%9C%89%E5%93%AA%E4%BA%9B
2.4 Dynamic Architecturehttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#24-dynamic-architecture
2.4.1 什么是 Dynamic Architecture？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#241-%E4%BB%80%E4%B9%88%E6%98%AF-dynamic-architecture
2.4.2 常用方法有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#242-%E5%B8%B8%E7%94%A8%E6%96%B9%E6%B3%95%E6%9C%89%E5%93%AA%E4%BA%9B
2.4.3 什么是 Lattice LSTM ，存在什么问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#243-%E4%BB%80%E4%B9%88%E6%98%AF-lattice-lstm-%E5%AD%98%E5%9C%A8%E4%BB%80%E4%B9%88%E9%97%AE%E9%A2%98
2.4.4 什么是 FLAT ，存在什么问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#244-%E4%BB%80%E4%B9%88%E6%98%AF-flat-%E5%AD%98%E5%9C%A8%E4%BB%80%E4%B9%88%E9%97%AE%E9%A2%98
2.5 Adaptive Embedding 范式https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#25-adaptive-embedding-%E8%8C%83%E5%BC%8F
2.5.1 什么是 Adaptive Embedding 范式？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#251-%E4%BB%80%E4%B9%88%E6%98%AF-adaptive-embedding-%E8%8C%83%E5%BC%8F
2.5.2 常用方法有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#252-%E5%B8%B8%E7%94%A8%E6%96%B9%E6%B3%95%E6%9C%89%E5%93%AA%E4%BA%9B
2.5.3 什么是 WC-LSTM ，存在什么问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#253-%E4%BB%80%E4%B9%88%E6%98%AF-wc-lstm-%E5%AD%98%E5%9C%A8%E4%BB%80%E4%B9%88%E9%97%AE%E9%A2%98
三、词汇/实体类型信息增强篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#%E4%B8%89%E8%AF%8D%E6%B1%87%E5%AE%9E%E4%BD%93%E7%B1%BB%E5%9E%8B%E4%BF%A1%E6%81%AF%E5%A2%9E%E5%BC%BA%E7%AF%87
3.1 什么是 词汇/实体类型信息增强？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#31-%E4%BB%80%E4%B9%88%E6%98%AF-%E8%AF%8D%E6%B1%87%E5%AE%9E%E4%BD%93%E7%B1%BB%E5%9E%8B%E4%BF%A1%E6%81%AF%E5%A2%9E%E5%BC%BA
3.2 为什么说 「词汇/实体类型信息增强」 方法对于中文 NER 任务有效呢？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#32-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%AF%B4-%E8%AF%8D%E6%B1%87%E5%AE%9E%E4%BD%93%E7%B1%BB%E5%9E%8B%E4%BF%A1%E6%81%AF%E5%A2%9E%E5%BC%BA-%E6%96%B9%E6%B3%95%E5%AF%B9%E4%BA%8E%E4%B8%AD%E6%96%87-ner-%E4%BB%BB%E5%8A%A1%E6%9C%89%E6%95%88%E5%91%A2
3.3 词汇/实体类型信息增强 方法有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#33-%E8%AF%8D%E6%B1%87%E5%AE%9E%E4%BD%93%E7%B1%BB%E5%9E%8B%E4%BF%A1%E6%81%AF%E5%A2%9E%E5%BC%BA-%E6%96%B9%E6%B3%95%E6%9C%89%E5%93%AA%E4%BA%9B
3.4 什么是 LEX-BERT ？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/ChineseNer/readme.md#34-%E4%BB%80%E4%B9%88%E6%98%AF-lex-bert-
【关于 命名实体识别 trick 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/NERtrick.md
trick 1：领域词典匹配https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#trick-1%E9%A2%86%E5%9F%9F%E8%AF%8D%E5%85%B8%E5%8C%B9%E9%85%8D
trick 2：规则抽取https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#trick-2%E8%A7%84%E5%88%99%E6%8A%BD%E5%8F%96
trick 3：词向量选取：词向量 or 字向量？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#trick-3%E8%AF%8D%E5%90%91%E9%87%8F%E9%80%89%E5%8F%96%E8%AF%8D%E5%90%91%E9%87%8F-or-%E5%AD%97%E5%90%91%E9%87%8F
trick 4：特征提取器 如何选择？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#trick-4%E7%89%B9%E5%BE%81%E6%8F%90%E5%8F%96%E5%99%A8-%E5%A6%82%E4%BD%95%E9%80%89%E6%8B%A9
trick 5：专有名称 怎么 处理？【注：这一点来自于 命名实体识别的几点心得 】https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#trick-5%E4%B8%93%E6%9C%89%E5%90%8D%E7%A7%B0-%E6%80%8E%E4%B9%88-%E5%A4%84%E7%90%86%E6%B3%A8%E8%BF%99%E4%B8%80%E7%82%B9%E6%9D%A5%E8%87%AA%E4%BA%8E-%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E7%9A%84%E5%87%A0%E7%82%B9%E5%BF%83%E5%BE%97-
trick 6：标注数据 不足怎么处理？【这个问题可以说是现在很多小厂最头疼的问题】https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#trick-6%E6%A0%87%E6%B3%A8%E6%95%B0%E6%8D%AE-%E4%B8%8D%E8%B6%B3%E6%80%8E%E4%B9%88%E5%A4%84%E7%90%86%E8%BF%99%E4%B8%AA%E9%97%AE%E9%A2%98%E5%8F%AF%E4%BB%A5%E8%AF%B4%E6%98%AF%E7%8E%B0%E5%9C%A8%E5%BE%88%E5%A4%9A%E5%B0%8F%E5%8E%82%E6%9C%80%E5%A4%B4%E7%96%BC%E7%9A%84%E9%97%AE%E9%A2%98
trick 7：嵌套命名实体识别怎么处理 【注：参考 资料3】https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#trick-7%E5%B5%8C%E5%A5%97%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E6%80%8E%E4%B9%88%E5%A4%84%E7%90%86-%E6%B3%A8%E5%8F%82%E8%80%83-%E8%B5%84%E6%96%993
7.1 什么是实体嵌套？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#71-%E4%BB%80%E4%B9%88%E6%98%AF%E5%AE%9E%E4%BD%93%E5%B5%8C%E5%A5%97
7.2 与 传统命名实体识别任务的区别https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#72-%E4%B8%8E-%E4%BC%A0%E7%BB%9F%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E4%BB%BB%E5%8A%A1%E7%9A%84%E5%8C%BA%E5%88%AB
7.3 解决方法：https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#73-%E8%A7%A3%E5%86%B3%E6%96%B9%E6%B3%95
7.3.1 方法一：序列标注https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#731-%E6%96%B9%E6%B3%95%E4%B8%80%E5%BA%8F%E5%88%97%E6%A0%87%E6%B3%A8
7.3.2 方法二：指针标注https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#732-%E6%96%B9%E6%B3%95%E4%BA%8C%E6%8C%87%E9%92%88%E6%A0%87%E6%B3%A8
7.3.3 方法三：多头标注https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#733-%E6%96%B9%E6%B3%95%E4%B8%89%E5%A4%9A%E5%A4%B4%E6%A0%87%E6%B3%A8
7.3.4 方法四：片段排列https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#734-%E6%96%B9%E6%B3%95%E5%9B%9B%E7%89%87%E6%AE%B5%E6%8E%92%E5%88%97
trick 8：为什么说 「词汇增强」 方法对于中文 NER 任务有效？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#trick-8%E4%B8%BA%E4%BB%80%E4%B9%88%E8%AF%B4-%E8%AF%8D%E6%B1%87%E5%A2%9E%E5%BC%BA-%E6%96%B9%E6%B3%95%E5%AF%B9%E4%BA%8E%E4%B8%AD%E6%96%87-ner-%E4%BB%BB%E5%8A%A1%E6%9C%89%E6%95%88
trick 9：NER实体span过长怎么办？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#trick-9ner%E5%AE%9E%E4%BD%93span%E8%BF%87%E9%95%BF%E6%80%8E%E4%B9%88%E5%8A%9E
trick 10: NER 标注数据噪声问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#trick-10-ner-%E6%A0%87%E6%B3%A8%E6%95%B0%E6%8D%AE%E5%99%AA%E5%A3%B0%E9%97%AE%E9%A2%98
trick 11： 给定两个命名实体识别任务，一个任务数据量足够，另外一个数据量很少，可以怎么做？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#trick-11-%E7%BB%99%E5%AE%9A%E4%B8%A4%E4%B8%AA%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E4%BB%BB%E5%8A%A1%E4%B8%80%E4%B8%AA%E4%BB%BB%E5%8A%A1%E6%95%B0%E6%8D%AE%E9%87%8F%E8%B6%B3%E5%A4%9F%E5%8F%A6%E5%A4%96%E4%B8%80%E4%B8%AA%E6%95%B0%E6%8D%AE%E9%87%8F%E5%BE%88%E5%B0%91%E5%8F%AF%E4%BB%A5%E6%80%8E%E4%B9%88%E5%81%9A
trick 12： NER 标注数据不均衡问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/ner/NERtrick/readme.md#trick-12-ner-%E6%A0%87%E6%B3%A8%E6%95%B0%E6%8D%AE%E4%B8%8D%E5%9D%87%E8%A1%A1%E9%97%AE%E9%A2%98
【关于 关系抽取】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction
https://github.com/turned123/NLP-Interview-Notes#412-关于-关系抽取那些你不知道的事
【关于 关系抽取】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 什么是关系抽取？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#11-%E4%BB%80%E4%B9%88%E6%98%AF%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96
1.2 关系抽取技术有哪些类型？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#12-%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E6%8A%80%E6%9C%AF%E6%9C%89%E5%93%AA%E4%BA%9B%E7%B1%BB%E5%9E%8B
1.3 常见的关系抽取流程是怎么做的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#13-%E5%B8%B8%E8%A7%81%E7%9A%84%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E6%B5%81%E7%A8%8B%E6%98%AF%E6%80%8E%E4%B9%88%E5%81%9A%E7%9A%84
二、经典关系抽取篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#%E4%BA%8C%E7%BB%8F%E5%85%B8%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E7%AF%87
2.1 模板匹配方法是指什么？有什么优缺点？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#21-%E6%A8%A1%E6%9D%BF%E5%8C%B9%E9%85%8D%E6%96%B9%E6%B3%95%E6%98%AF%E6%8C%87%E4%BB%80%E4%B9%88%E6%9C%89%E4%BB%80%E4%B9%88%E4%BC%98%E7%BC%BA%E7%82%B9
2.2 远监督关系抽取是指什么？它有什么优缺点？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#22-%E8%BF%9C%E7%9B%91%E7%9D%A3%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E6%98%AF%E6%8C%87%E4%BB%80%E4%B9%88%E5%AE%83%E6%9C%89%E4%BB%80%E4%B9%88%E4%BC%98%E7%BC%BA%E7%82%B9
2.3 什么是关系重叠？复杂关系问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#23-%E4%BB%80%E4%B9%88%E6%98%AF%E5%85%B3%E7%B3%BB%E9%87%8D%E5%8F%A0%E5%A4%8D%E6%9D%82%E5%85%B3%E7%B3%BB%E9%97%AE%E9%A2%98
2.4 联合抽取是什么？难点在哪里？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#24-%E8%81%94%E5%90%88%E6%8A%BD%E5%8F%96%E6%98%AF%E4%BB%80%E4%B9%88%E9%9A%BE%E7%82%B9%E5%9C%A8%E5%93%AA%E9%87%8C
2.5 联合抽取总体上有哪些方法？各有哪些缺点？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#25-%E8%81%94%E5%90%88%E6%8A%BD%E5%8F%96%E6%80%BB%E4%BD%93%E4%B8%8A%E6%9C%89%E5%93%AA%E4%BA%9B%E6%96%B9%E6%B3%95%E5%90%84%E6%9C%89%E5%93%AA%E4%BA%9B%E7%BC%BA%E7%82%B9
2.6 介绍基于共享参数的联合抽取方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#26-%E4%BB%8B%E7%BB%8D%E5%9F%BA%E4%BA%8E%E5%85%B1%E4%BA%AB%E5%8F%82%E6%95%B0%E7%9A%84%E8%81%94%E5%90%88%E6%8A%BD%E5%8F%96%E6%96%B9%E6%B3%95
依存结构树：End-to-End Relation Extraction using LSTMs on Sequences and Tree Structureshttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#%E4%BE%9D%E5%AD%98%E7%BB%93%E6%9E%84%E6%A0%91end-to-end-relation-extraction-using-lstms-on-sequences-and-tree-structures
指针网络，Going out on a limb: Joint Extraction of Entity Mentions and Relations without Dependency Treeshttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#%E6%8C%87%E9%92%88%E7%BD%91%E7%BB%9Cgoing-out-on-a-limb-joint-extraction-of-entity-mentions-and-relations-without-dependency-trees
Copy机制+seq2seq：Extracting Relational Facts by an End-to-End Neural Model with Copy Mechanism[19]https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#copy%E6%9C%BA%E5%88%B6seq2seqextracting-relational-facts-by-an-end-to-end-neural-model-with-copy-mechanism19
多头选择机制+sigmoid：Joint entity recognition and relation extraction as a multi-head selection problemhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#%E5%A4%9A%E5%A4%B4%E9%80%89%E6%8B%A9%E6%9C%BA%E5%88%B6sigmoidjoint-entity-recognition-and-relation-extraction-as-a-multi-head-selection-problem
SPO问题+指针网络，Joint Extraction of Entities and Relations Based on a Novel Decomposition Strategyhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#spo%E9%97%AE%E9%A2%98%E6%8C%87%E9%92%88%E7%BD%91%E7%BB%9Cjoint-extraction-of-entities-and-relations-based-on-a-novel-decomposition-strategy
多轮对话+强化学习 ：Entity-Relation Extraction as Multi-Turn Question Answeringhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#%E5%A4%9A%E8%BD%AE%E5%AF%B9%E8%AF%9D%E5%BC%BA%E5%8C%96%E5%AD%A6%E4%B9%A0-entity-relation-extraction-as-multi-turn-question-answering
输入端的片段排列： Span-Level Model for Relation Extractionhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#%E8%BE%93%E5%85%A5%E7%AB%AF%E7%9A%84%E7%89%87%E6%AE%B5%E6%8E%92%E5%88%97-span-level-model-for-relation-extraction
输出端的片段排列：SpERT：Span-based Joint Entity and Relation Extraction with Transformer Pre-traininghttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#%E8%BE%93%E5%87%BA%E7%AB%AF%E7%9A%84%E7%89%87%E6%AE%B5%E6%8E%92%E5%88%97spertspan-based-joint-entity-and-relation-extraction-with-transformer-pre-training
2.7 介绍基于联合解码的联合抽取方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#27-%E4%BB%8B%E7%BB%8D%E5%9F%BA%E4%BA%8E%E8%81%94%E5%90%88%E8%A7%A3%E7%A0%81%E7%9A%84%E8%81%94%E5%90%88%E6%8A%BD%E5%8F%96%E6%96%B9%E6%B3%95
Joint extraction of entities and relations based on a novel tagging schemehttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#joint-extraction-of-entities-and-relations-based-on-a-novel-tagging-scheme
Joint Extraction of Entities and Overlapping Relations Using Position-Attentive Sequence Labelinghttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#joint-extraction-of-entities-and-overlapping-relations-using-position-attentive-sequence-labeling
Joint extraction of entities and relations based on a novel tagging schemehttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#joint-extraction-of-entities-and-relations-based-on-a-novel-tagging-scheme-1
2.8 实体关系抽取的前沿技术和挑战有哪些？如何解决低资源和复杂样本下的实体关系抽取？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#28-%E5%AE%9E%E4%BD%93%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E7%9A%84%E5%89%8D%E6%B2%BF%E6%8A%80%E6%9C%AF%E5%92%8C%E6%8C%91%E6%88%98%E6%9C%89%E5%93%AA%E4%BA%9B%E5%A6%82%E4%BD%95%E8%A7%A3%E5%86%B3%E4%BD%8E%E8%B5%84%E6%BA%90%E5%92%8C%E5%A4%8D%E6%9D%82%E6%A0%B7%E6%9C%AC%E4%B8%8B%E7%9A%84%E5%AE%9E%E4%BD%93%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96
三、文档级关系抽取篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#%E4%B8%89%E6%96%87%E6%A1%A3%E7%BA%A7%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E7%AF%87
3.1 文档级关系抽取与经典关系抽取有何区别？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#31-%E6%96%87%E6%A1%A3%E7%BA%A7%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E4%B8%8E%E7%BB%8F%E5%85%B8%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E6%9C%89%E4%BD%95%E5%8C%BA%E5%88%AB
3.2 文档级别关系抽取中面临什么样的问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#32-%E6%96%87%E6%A1%A3%E7%BA%A7%E5%88%AB%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E4%B8%AD%E9%9D%A2%E4%B8%B4%E4%BB%80%E4%B9%88%E6%A0%B7%E7%9A%84%E9%97%AE%E9%A2%98
3.3 文档级关系抽取的方法有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#33-%E6%96%87%E6%A1%A3%E7%BA%A7%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E7%9A%84%E6%96%B9%E6%B3%95%E6%9C%89%E5%93%AA%E4%BA%9B
3.3.1 基于BERT-like的文档关系抽取是怎么做的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#331-%E5%9F%BA%E4%BA%8Ebert-like%E7%9A%84%E6%96%87%E6%A1%A3%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E6%98%AF%E6%80%8E%E4%B9%88%E5%81%9A%E7%9A%84
3.3.2 基于graph的文档关系抽取是怎么做的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#332-%E5%9F%BA%E4%BA%8Egraph%E7%9A%84%E6%96%87%E6%A1%A3%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E6%98%AF%E6%80%8E%E4%B9%88%E5%81%9A%E7%9A%84
3.4 文档级关系抽取常见数据集有哪些以及其评估方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/RelationExtraction/readme.md#34-%E6%96%87%E6%A1%A3%E7%BA%A7%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E5%B8%B8%E8%A7%81%E6%95%B0%E6%8D%AE%E9%9B%86%E6%9C%89%E5%93%AA%E4%BA%9B%E4%BB%A5%E5%8F%8A%E5%85%B6%E8%AF%84%E4%BC%B0%E6%96%B9%E6%B3%95
【关于 事件抽取】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction
https://github.com/turned123/NLP-Interview-Notes#413-关于-事件抽取那些你不知道的事
【关于 事件抽取】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction
一、原理篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#%E4%B8%80%E5%8E%9F%E7%90%86%E7%AF%87
1.1 什么是事件？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#11-%E4%BB%80%E4%B9%88%E6%98%AF%E4%BA%8B%E4%BB%B6
1.2 什么是事件抽取？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#12-%E4%BB%80%E4%B9%88%E6%98%AF%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96
1.3 ACE测评中事件抽取涉及的几个基本术语及任务是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#13-ace%E6%B5%8B%E8%AF%84%E4%B8%AD%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E6%B6%89%E5%8F%8A%E7%9A%84%E5%87%A0%E4%B8%AA%E5%9F%BA%E6%9C%AC%E6%9C%AF%E8%AF%AD%E5%8F%8A%E4%BB%BB%E5%8A%A1%E6%98%AF%E4%BB%80%E4%B9%88
1.4 事件抽取怎么发展的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#14-%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E6%80%8E%E4%B9%88%E5%8F%91%E5%B1%95%E7%9A%84
1.5 事件抽取存在什么问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#15-%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E5%AD%98%E5%9C%A8%E4%BB%80%E4%B9%88%E9%97%AE%E9%A2%98
二、基本任务篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#%E4%BA%8C%E5%9F%BA%E6%9C%AC%E4%BB%BB%E5%8A%A1%E7%AF%87
2.1 触发词检测https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#21-%E8%A7%A6%E5%8F%91%E8%AF%8D%E6%A3%80%E6%B5%8B
2.1.1 什么是触发词检测？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#211-%E4%BB%80%E4%B9%88%E6%98%AF%E8%A7%A6%E5%8F%91%E8%AF%8D%E6%A3%80%E6%B5%8B
2.1.2 触发词检测有哪些方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#212-%E8%A7%A6%E5%8F%91%E8%AF%8D%E6%A3%80%E6%B5%8B%E6%9C%89%E5%93%AA%E4%BA%9B%E6%96%B9%E6%B3%95
2.2 类型识别https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#22-%E7%B1%BB%E5%9E%8B%E8%AF%86%E5%88%AB
2.2.1 什么是类型识别？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#221-%E4%BB%80%E4%B9%88%E6%98%AF%E7%B1%BB%E5%9E%8B%E8%AF%86%E5%88%AB
2.2.2 类型识别有哪些方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#222-%E7%B1%BB%E5%9E%8B%E8%AF%86%E5%88%AB%E6%9C%89%E5%93%AA%E4%BA%9B%E6%96%B9%E6%B3%95
2.3 角色识别https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#23-%E8%A7%92%E8%89%B2%E8%AF%86%E5%88%AB
2.3.1 什么是角色识别？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#231-%E4%BB%80%E4%B9%88%E6%98%AF%E8%A7%92%E8%89%B2%E8%AF%86%E5%88%AB
2.3.2 角色识别有哪些方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#232-%E8%A7%92%E8%89%B2%E8%AF%86%E5%88%AB%E6%9C%89%E5%93%AA%E4%BA%9B%E6%96%B9%E6%B3%95
2.4 论元检测https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#24-%E8%AE%BA%E5%85%83%E6%A3%80%E6%B5%8B
2.4.1 什么是论元检测？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#241-%E4%BB%80%E4%B9%88%E6%98%AF%E8%AE%BA%E5%85%83%E6%A3%80%E6%B5%8B
2.4.2 论元检测有哪些方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#242-%E8%AE%BA%E5%85%83%E6%A3%80%E6%B5%8B%E6%9C%89%E5%93%AA%E4%BA%9B%E6%96%B9%E6%B3%95
三、常见方法篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#%E4%B8%89%E5%B8%B8%E8%A7%81%E6%96%B9%E6%B3%95%E7%AF%87
3.1 模式匹配方法怎么用在事件抽取中?https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#31-%E6%A8%A1%E5%BC%8F%E5%8C%B9%E9%85%8D%E6%96%B9%E6%B3%95%E6%80%8E%E4%B9%88%E7%94%A8%E5%9C%A8%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E4%B8%AD
3.2 统计机器学习方法怎么用在事件抽取中?https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#32-%E7%BB%9F%E8%AE%A1%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E6%96%B9%E6%B3%95%E6%80%8E%E4%B9%88%E7%94%A8%E5%9C%A8%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E4%B8%AD
3.3 深度学习方法怎么用在事件抽取中?https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#33-%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E6%96%B9%E6%B3%95%E6%80%8E%E4%B9%88%E7%94%A8%E5%9C%A8%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E4%B8%AD
四、数据集及评价指标篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#%E5%9B%9B%E6%95%B0%E6%8D%AE%E9%9B%86%E5%8F%8A%E8%AF%84%E4%BB%B7%E6%8C%87%E6%A0%87%E7%AF%87
4.1 事件抽取中常见的英文数据集有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#41-%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E4%B8%AD%E5%B8%B8%E8%A7%81%E7%9A%84%E8%8B%B1%E6%96%87%E6%95%B0%E6%8D%AE%E9%9B%86%E6%9C%89%E5%93%AA%E4%BA%9B
4.2 事件抽取中常见的中文数据集有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#42-%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E4%B8%AD%E5%B8%B8%E8%A7%81%E7%9A%84%E4%B8%AD%E6%96%87%E6%95%B0%E6%8D%AE%E9%9B%86%E6%9C%89%E5%93%AA%E4%BA%9B
4.3 事件抽取的评价指标是什么？怎么计算的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#43-%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E7%9A%84%E8%AF%84%E4%BB%B7%E6%8C%87%E6%A0%87%E6%98%AF%E4%BB%80%E4%B9%88%E6%80%8E%E4%B9%88%E8%AE%A1%E7%AE%97%E7%9A%84
五、对比篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#%E4%BA%94%E5%AF%B9%E6%AF%94%E7%AF%87
5.1 事件抽取和命名实体识别（即实体抽取）有什么异同？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#51-%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E5%92%8C%E5%91%BD%E5%90%8D%E5%AE%9E%E4%BD%93%E8%AF%86%E5%88%AB%E5%8D%B3%E5%AE%9E%E4%BD%93%E6%8A%BD%E5%8F%96%E6%9C%89%E4%BB%80%E4%B9%88%E5%BC%82%E5%90%8C
5.2 事件抽取和关系抽取有什么异同？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#52-%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E5%92%8C%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96%E6%9C%89%E4%BB%80%E4%B9%88%E5%BC%82%E5%90%8C
5.3 什么是事理图谱？有哪些事件关系类型？事理图谱怎么构建？主要技术领域及当前发展热点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#53-%E4%BB%80%E4%B9%88%E6%98%AF%E4%BA%8B%E7%90%86%E5%9B%BE%E8%B0%B1%E6%9C%89%E5%93%AA%E4%BA%9B%E4%BA%8B%E4%BB%B6%E5%85%B3%E7%B3%BB%E7%B1%BB%E5%9E%8B%E4%BA%8B%E7%90%86%E5%9B%BE%E8%B0%B1%E6%80%8E%E4%B9%88%E6%9E%84%E5%BB%BA%E4%B8%BB%E8%A6%81%E6%8A%80%E6%9C%AF%E9%A2%86%E5%9F%9F%E5%8F%8A%E5%BD%93%E5%89%8D%E5%8F%91%E5%B1%95%E7%83%AD%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
六、应用篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#%E5%85%AD%E5%BA%94%E7%94%A8%E7%AF%87
七、拓展篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#%E4%B8%83%E6%8B%93%E5%B1%95%E7%AF%87
7.1 事件抽取论文综述https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#71-%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E8%AE%BA%E6%96%87%E7%BB%BC%E8%BF%B0
7.2 事件抽取常见问题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/EventExtraction/readme.md#72-%E4%BA%8B%E4%BB%B6%E6%8A%BD%E5%8F%96%E5%B8%B8%E8%A7%81%E9%97%AE%E9%A2%98
【关于 NLP 预训练算法】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining
https://github.com/turned123/NLP-Interview-Notes#42-关于-nlp-预训练算法那些你不知道的事
【关于TF-idf】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/tfidf
一、one-hot 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#%E4%B8%80one-hot-%E7%AF%87
1.1 为什么有 one-hot ？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#11-%E4%B8%BA%E4%BB%80%E4%B9%88%E6%9C%89-one-hot-
1.2 one-hot 是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#12-one-hot-%E6%98%AF%E4%BB%80%E4%B9%88
1.3 one-hot 有什么特点?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#13-one-hot-%E6%9C%89%E4%BB%80%E4%B9%88%E7%89%B9%E7%82%B9
1.4 one-hot 存在哪些问题?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#14-one-hot-%E5%AD%98%E5%9C%A8%E5%93%AA%E4%BA%9B%E9%97%AE%E9%A2%98
二、TF-IDF 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#%E4%BA%8Ctf-idf-%E7%AF%87
2.1 什么是 TF-IDF？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#21-%E4%BB%80%E4%B9%88%E6%98%AF-tf-idf
2.2 TF-IDF 如何评估词的重要程度？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#22--tf-idf-%E5%A6%82%E4%BD%95%E8%AF%84%E4%BC%B0%E8%AF%8D%E7%9A%84%E9%87%8D%E8%A6%81%E7%A8%8B%E5%BA%A6
2.3 TF-IDF 的思想是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#23--tf-idf-%E7%9A%84%E6%80%9D%E6%83%B3%E6%98%AF%E4%BB%80%E4%B9%88
2.4 TF-IDF 的计算公式是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#24--tf-idf-%E7%9A%84%E8%AE%A1%E7%AE%97%E5%85%AC%E5%BC%8F%E6%98%AF%E4%BB%80%E4%B9%88
2.5 TF-IDF 怎么描述？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#25--tf-idf-%E6%80%8E%E4%B9%88%E6%8F%8F%E8%BF%B0
2.6 TF-IDF 的优点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#26--tf-idf-%E7%9A%84%E4%BC%98%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
2.7 TF-IDF 的缺点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#27--tf-idf-%E7%9A%84%E7%BC%BA%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
2.8 TF-IDF 的应用？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/tfidf/readme.md#28--tf-idf-%E7%9A%84%E5%BA%94%E7%94%A8
【关于word2vec】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/word2vec
一、Wordvec 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#%E4%B8%80wordvec-%E4%BB%8B%E7%BB%8D%E7%AF%87
1.1 Wordvec 指什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#11-wordvec-%E6%8C%87%E4%BB%80%E4%B9%88
1.2 Wordvec 中 CBOW 指什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#12-wordvec-%E4%B8%AD-cbow-%E6%8C%87%E4%BB%80%E4%B9%88
1.3 Wordvec 中 Skip-gram 指什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#13-wordvec-%E4%B8%AD-skip-gram-%E6%8C%87%E4%BB%80%E4%B9%88
1.4 CBOW vs Skip-gram 哪一个好？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#14-cbow-vs-skip-gram-%E5%93%AA%E4%B8%80%E4%B8%AA%E5%A5%BD
二、Wordvec 优化篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#%E4%BA%8Cwordvec-%E4%BC%98%E5%8C%96%E7%AF%87
2.1 Word2vec 中 霍夫曼树 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#21--word2vec-%E4%B8%AD-%E9%9C%8D%E5%A4%AB%E6%9B%BC%E6%A0%91-%E6%98%AF%E4%BB%80%E4%B9%88
2.2 Word2vec 中 为什么要使用 霍夫曼树？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#22--word2vec-%E4%B8%AD-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E4%BD%BF%E7%94%A8-%E9%9C%8D%E5%A4%AB%E6%9B%BC%E6%A0%91
2.3 Word2vec 中使用 霍夫曼树 的好处？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#23--word2vec-%E4%B8%AD%E4%BD%BF%E7%94%A8-%E9%9C%8D%E5%A4%AB%E6%9B%BC%E6%A0%91-%E7%9A%84%E5%A5%BD%E5%A4%84
2.4 为什么 Word2vec 中会用到 负采样？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#24-%E4%B8%BA%E4%BB%80%E4%B9%88-word2vec-%E4%B8%AD%E4%BC%9A%E7%94%A8%E5%88%B0-%E8%B4%9F%E9%87%87%E6%A0%B7
2.5 Word2vec 中会用到 负采样 是什么样？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#25-word2vec-%E4%B8%AD%E4%BC%9A%E7%94%A8%E5%88%B0-%E8%B4%9F%E9%87%87%E6%A0%B7-%E6%98%AF%E4%BB%80%E4%B9%88%E6%A0%B7
2.6 Word2vec 中 负采样 的采样方式？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#26--word2vec-%E4%B8%AD-%E8%B4%9F%E9%87%87%E6%A0%B7-%E7%9A%84%E9%87%87%E6%A0%B7%E6%96%B9%E5%BC%8F
三、Wordvec 对比篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#%E4%B8%89wordvec-%E5%AF%B9%E6%AF%94%E7%AF%87
3.1 word2vec和NNLM对比有什么区别？（word2vec vs NNLM）https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#31-word2vec%E5%92%8Cnnlm%E5%AF%B9%E6%AF%94%E6%9C%89%E4%BB%80%E4%B9%88%E5%8C%BA%E5%88%ABword2vec-vs-nnlm
3.2 word2vec和tf-idf 在相似度计算时的区别？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#32-word2vec%E5%92%8Ctf-idf-%E5%9C%A8%E7%9B%B8%E4%BC%BC%E5%BA%A6%E8%AE%A1%E7%AE%97%E6%97%B6%E7%9A%84%E5%8C%BA%E5%88%AB
四、word2vec 实战篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#%E5%9B%9Bword2vec-%E5%AE%9E%E6%88%98%E7%AF%87
4.1 word2vec训练trick，window设置多大？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#41-word2vec%E8%AE%AD%E7%BB%83trickwindow%E8%AE%BE%E7%BD%AE%E5%A4%9A%E5%A4%A7
4.1 word2vec训练trick，词向量纬度，大与小有什么影响，还有其他参数？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/word2vec/readme.md#41-word2vec%E8%AE%AD%E7%BB%83trick%E8%AF%8D%E5%90%91%E9%87%8F%E7%BA%AC%E5%BA%A6%E5%A4%A7%E4%B8%8E%E5%B0%8F%E6%9C%89%E4%BB%80%E4%B9%88%E5%BD%B1%E5%93%8D%E8%BF%98%E6%9C%89%E5%85%B6%E4%BB%96%E5%8F%82%E6%95%B0
【关于FastText】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/fasttext
一、fastText 动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#%E4%B8%80fasttext--%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 word-level Model 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#11-word-level-model-%E6%98%AF%E4%BB%80%E4%B9%88
1.2 word-level Model 存在什么问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#12-word-level-model-%E5%AD%98%E5%9C%A8%E4%BB%80%E4%B9%88%E9%97%AE%E9%A2%98
1.3 Character-Level Model 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#13-character-level-model-%E6%98%AF%E4%BB%80%E4%B9%88
1.4 Character-Level Model 优点？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#14-character-level-model-%E4%BC%98%E7%82%B9
1.5 Character-Level Model 存在问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#15-character-level-model-%E5%AD%98%E5%9C%A8%E9%97%AE%E9%A2%98
1.6 Character-Level Model 问题的解决方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#16-character-level-model-%E9%97%AE%E9%A2%98%E7%9A%84%E8%A7%A3%E5%86%B3%E6%96%B9%E6%B3%95
二、 词内的n-gram信息(subword n-gram information) 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#%E4%BA%8C-%E8%AF%8D%E5%86%85%E7%9A%84n-gram%E4%BF%A1%E6%81%AFsubword-n-gram-information-%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1 引言https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#21-%E5%BC%95%E8%A8%80
2.2 fastText 是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#22-fasttext-%E6%98%AF%E4%BB%80%E4%B9%88
2.3 fastText 的结构是什么样?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#23-fasttext-%E7%9A%84%E7%BB%93%E6%9E%84%E6%98%AF%E4%BB%80%E4%B9%88%E6%A0%B7
2.4 为什么 fastText 要使用词内的n-gram信息(subword n-gram information)?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#24-%E4%B8%BA%E4%BB%80%E4%B9%88-fasttext-%E8%A6%81%E4%BD%BF%E7%94%A8%E8%AF%8D%E5%86%85%E7%9A%84n-gram%E4%BF%A1%E6%81%AFsubword-n-gram-information
2.5 fastText 词内的n-gram信息(subword n-gram information) 介绍?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#25-fasttext-%E8%AF%8D%E5%86%85%E7%9A%84n-gram%E4%BF%A1%E6%81%AFsubword-n-gram-information-%E4%BB%8B%E7%BB%8D
2.6 fastText 词内的n-gram信息 的 训练过程?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#26-fasttext-%E8%AF%8D%E5%86%85%E7%9A%84n-gram%E4%BF%A1%E6%81%AF-%E7%9A%84-%E8%AE%AD%E7%BB%83%E8%BF%87%E7%A8%8B
2.7 fastText 词内的n-gram信息 存在问题?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#27-fasttext-%E8%AF%8D%E5%86%85%E7%9A%84n-gram%E4%BF%A1%E6%81%AF-%E5%AD%98%E5%9C%A8%E9%97%AE%E9%A2%98
三、 层次化Softmax回归(Hierarchical Softmax) 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#%E4%B8%89-%E5%B1%82%E6%AC%A1%E5%8C%96softmax%E5%9B%9E%E5%BD%92hierarchical-softmax-%E4%BB%8B%E7%BB%8D%E7%AF%87
3.1 为什么要用 层次化Softmax回归(Hierarchical Softmax) ？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#31-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E7%94%A8-%E5%B1%82%E6%AC%A1%E5%8C%96softmax%E5%9B%9E%E5%BD%92hierarchical-softmax-
3.2 层次化Softmax回归(Hierarchical Softmax) 的思想是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#32-%E5%B1%82%E6%AC%A1%E5%8C%96softmax%E5%9B%9E%E5%BD%92hierarchical-softmax-%E7%9A%84%E6%80%9D%E6%83%B3%E6%98%AF%E4%BB%80%E4%B9%88
3.3 层次化Softmax回归(Hierarchical Softmax) 的步骤？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#33-%E5%B1%82%E6%AC%A1%E5%8C%96softmax%E5%9B%9E%E5%BD%92hierarchical-softmax-%E7%9A%84%E6%AD%A5%E9%AA%A4
四、fastText 存在问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/fasttext/readme.md#%E5%9B%9Bfasttext-%E5%AD%98%E5%9C%A8%E9%97%AE%E9%A2%98
【关于Elmo】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/elmo
一、Elmo 动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/elmo/readme.md#%E4%B8%80elmo-%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 为什么会有 Elmo？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/elmo/readme.md#11-%E4%B8%BA%E4%BB%80%E4%B9%88%E4%BC%9A%E6%9C%89-elmo
二、Elmo 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/elmo/readme.md#%E4%BA%8Celmo-%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1 Elmo 的 特点？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/elmo/readme.md#21-elmo-%E7%9A%84-%E7%89%B9%E7%82%B9
2.2 Elmo 的 思想是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/elmo/readme.md#22-elmo-%E7%9A%84-%E6%80%9D%E6%83%B3%E6%98%AF%E4%BB%80%E4%B9%88
三、Elmo 问题篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/elmo/readme.md#%E4%B8%89elmo-%E9%97%AE%E9%A2%98%E7%AF%87
3.1 Elmo 存在的问题是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/elmo/readme.md#31-elmo-%E5%AD%98%E5%9C%A8%E7%9A%84%E9%97%AE%E9%A2%98%E6%98%AF%E4%BB%80%E4%B9%88
【关于Bert】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert
【关于Bert】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/readme.md
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 【演变史】one-hot 存在问题?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#11-%E6%BC%94%E5%8F%98%E5%8F%B2one-hot-%E5%AD%98%E5%9C%A8%E9%97%AE%E9%A2%98
1.2【演变史】wordvec 存在问题?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#12%E6%BC%94%E5%8F%98%E5%8F%B2wordvec-%E5%AD%98%E5%9C%A8%E9%97%AE%E9%A2%98
1.3【演变史】fastText 存在问题?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#13%E6%BC%94%E5%8F%98%E5%8F%B2fasttext-%E5%AD%98%E5%9C%A8%E9%97%AE%E9%A2%98
1.4【演变史】elmo 存在问题?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#14%E6%BC%94%E5%8F%98%E5%8F%B2elmo-%E5%AD%98%E5%9C%A8%E9%97%AE%E9%A2%98
二、Bert 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#%E4%BA%8Cbert-%E7%AF%87
2.1 Bert 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#21-bert-%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1.1【BERT】Bert 是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#211bertbert-%E6%98%AF%E4%BB%80%E4%B9%88
2.1.2【BERT】Bert 三个关键点？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#212bertbert-%E4%B8%89%E4%B8%AA%E5%85%B3%E9%94%AE%E7%82%B9
2.2 Bert 输入输出表征篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#22-bert-%E8%BE%93%E5%85%A5%E8%BE%93%E5%87%BA%E8%A1%A8%E5%BE%81%E7%AF%87
2.2.1 【BERT】Bert 输入输出表征长啥样？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#221-bertbert-%E8%BE%93%E5%85%A5%E8%BE%93%E5%87%BA%E8%A1%A8%E5%BE%81%E9%95%BF%E5%95%A5%E6%A0%B7
2.3 【BERT】Bert 预训练篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#23-bertbert-%E9%A2%84%E8%AE%AD%E7%BB%83%E7%AF%87
2.3.1 【BERT】Bert 预训练任务介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#231-bertbert-%E9%A2%84%E8%AE%AD%E7%BB%83%E4%BB%BB%E5%8A%A1%E4%BB%8B%E7%BB%8D
2.3.2 【BERT】Bert 预训练任务 之 Masked LM 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#232-bertbert-%E9%A2%84%E8%AE%AD%E7%BB%83%E4%BB%BB%E5%8A%A1-%E4%B9%8B-masked-lm-%E7%AF%87
2.3.2.1 【BERT】 Bert 为什么需要预训练任务 Masked LM ？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#2321-bert-bert-%E4%B8%BA%E4%BB%80%E4%B9%88%E9%9C%80%E8%A6%81%E9%A2%84%E8%AE%AD%E7%BB%83%E4%BB%BB%E5%8A%A1-masked-lm-
2.3.2.2 【BERT】 Bert 预训练任务 Masked LM 怎么做？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#2322-bert-bert-%E9%A2%84%E8%AE%AD%E7%BB%83%E4%BB%BB%E5%8A%A1-masked-lm-%E6%80%8E%E4%B9%88%E5%81%9A
2.3.2.3 【BERT】 Bert 预训练任务 Masked LM 存在问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#2323-bert-bert-%E9%A2%84%E8%AE%AD%E7%BB%83%E4%BB%BB%E5%8A%A1-masked-lm-%E5%AD%98%E5%9C%A8%E9%97%AE%E9%A2%98
2.3.2.4 【BERT】 预训练和微调之间的不匹配的解决方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#2324-bert-%E9%A2%84%E8%AE%AD%E7%BB%83%E5%92%8C%E5%BE%AE%E8%B0%83%E4%B9%8B%E9%97%B4%E7%9A%84%E4%B8%8D%E5%8C%B9%E9%85%8D%E7%9A%84%E8%A7%A3%E5%86%B3%E6%96%B9%E6%B3%95
2.3.3 【BERT】Bert 预训练任务 之 Next Sentence Prediction 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#233-bertbert-%E9%A2%84%E8%AE%AD%E7%BB%83%E4%BB%BB%E5%8A%A1-%E4%B9%8B-next-sentence-prediction-%E7%AF%87
2.3.3.1 【BERT】Bert 为什么需要预训练任务 Next Sentence Prediction ？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#2331-bertbert-%E4%B8%BA%E4%BB%80%E4%B9%88%E9%9C%80%E8%A6%81%E9%A2%84%E8%AE%AD%E7%BB%83%E4%BB%BB%E5%8A%A1-next-sentence-prediction-
2.3.3.2 【BERT】 Bert 预训练任务 Next Sentence Prediction 怎么做？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#2332-bert-bert-%E9%A2%84%E8%AE%AD%E7%BB%83%E4%BB%BB%E5%8A%A1-next-sentence-prediction-%E6%80%8E%E4%B9%88%E5%81%9A
2.4 【BERT】 fine-turning 篇？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#24-bert-fine-turning-%E7%AF%87
2.4.1 【BERT】为什么 Bert 需要 fine-turning？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#241-bert%E4%B8%BA%E4%BB%80%E4%B9%88-bert-%E9%9C%80%E8%A6%81-fine-turning
2.4.2 【BERT】 Bert 如何 fine-turning？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#242-bert-bert-%E5%A6%82%E4%BD%95-fine-turning
2.5 【BERT】 Bert 损失函数篇？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#25-bert-bert-%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0%E7%AF%87
2.5.1 【BERT】BERT的两个预训练任务对应的损失函数是什么(用公式形式展示)？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#251-bertbert%E7%9A%84%E4%B8%A4%E4%B8%AA%E9%A2%84%E8%AE%AD%E7%BB%83%E4%BB%BB%E5%8A%A1%E5%AF%B9%E5%BA%94%E7%9A%84%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0%E6%98%AF%E4%BB%80%E4%B9%88%E7%94%A8%E5%85%AC%E5%BC%8F%E5%BD%A2%E5%BC%8F%E5%B1%95%E7%A4%BA
三、 对比篇？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#%E4%B8%89-%E5%AF%B9%E6%AF%94%E7%AF%87
3.1 【对比】多义词问题是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#31-%E5%AF%B9%E6%AF%94%E5%A4%9A%E4%B9%89%E8%AF%8D%E9%97%AE%E9%A2%98%E6%98%AF%E4%BB%80%E4%B9%88
3.2 【对比】word2vec 为什么解决不了多义词问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#32-%E5%AF%B9%E6%AF%94word2vec-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A7%A3%E5%86%B3%E4%B8%8D%E4%BA%86%E5%A4%9A%E4%B9%89%E8%AF%8D%E9%97%AE%E9%A2%98
3.3 【对比】GPT和BERT有什么不同？https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#33-%E5%AF%B9%E6%AF%94gpt%E5%92%8Cbert%E6%9C%89%E4%BB%80%E4%B9%88%E4%B8%8D%E5%90%8C
3.4 【对比】为什么 elmo、GPT、Bert能够解决多义词问题？（以 elmo 为例）https://github.com/turned123/NLP-Interview-Notes/blob/main/PreTraining/bert/readme.md#34-%E5%AF%B9%E6%AF%94%E4%B8%BA%E4%BB%80%E4%B9%88-elmogptbert%E8%83%BD%E5%A4%9F%E8%A7%A3%E5%86%B3%E5%A4%9A%E4%B9%89%E8%AF%8D%E9%97%AE%E9%A2%98%E4%BB%A5-elmo-%E4%B8%BA%E4%BE%8B
【关于 Bert 源码解析I 之 主体篇】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md
一、动机https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#%E4%B8%80%E5%8A%A8%E6%9C%BA
二、本文框架https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#%E4%BA%8C%E6%9C%AC%E6%96%87%E6%A1%86%E6%9E%B6
三、前言https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#%E4%B8%89%E5%89%8D%E8%A8%80
四、配置类 BertConfighttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#%E5%9B%9B%E9%85%8D%E7%BD%AE%E7%B1%BB-bertconfig
五、获取 词向量 （Embedding_lookup）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#%E4%BA%94%E8%8E%B7%E5%8F%96-%E8%AF%8D%E5%90%91%E9%87%8F-embedding_lookup
六、词向量 的后处理 （embedding_postprocessor）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#%E5%85%AD%E8%AF%8D%E5%90%91%E9%87%8F-%E7%9A%84%E5%90%8E%E5%A4%84%E7%90%86-embedding_postprocessor
6.1 介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#61-%E4%BB%8B%E7%BB%8D
6.2 特点https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#62-%E7%89%B9%E7%82%B9
6.3 代码实现https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#63-%E4%BB%A3%E7%A0%81%E5%AE%9E%E7%8E%B0
七、创建 attention mask （attention_mask）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#%E4%B8%83%E5%88%9B%E5%BB%BA-attention-mask-attention_mask
7.1 作用https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#71-%E4%BD%9C%E7%94%A8
7.2 代码https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#72-%E4%BB%A3%E7%A0%81
八、注意力层（attention layer）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#%E5%85%AB%E6%B3%A8%E6%84%8F%E5%8A%9B%E5%B1%82attention-layer
8.1 自注意力层（self-attention）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#81-%E8%87%AA%E6%B3%A8%E6%84%8F%E5%8A%9B%E5%B1%82self-attention
8.1.1 动机https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#811-%E5%8A%A8%E6%9C%BA
8.1.2 传统 Attentionhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#812-%E4%BC%A0%E7%BB%9F-attention
8.1.3 核心思想https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#813-%E6%A0%B8%E5%BF%83%E6%80%9D%E6%83%B3
8.1.4 目的https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#814-%E7%9B%AE%E7%9A%84
8.1.5 公式https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#815-%E5%85%AC%E5%BC%8F
8.1.6 步骤https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#816-%E6%AD%A5%E9%AA%A4
8.2 多头自注意力 （Multi-Headed Attention）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#82-%E5%A4%9A%E5%A4%B4%E8%87%AA%E6%B3%A8%E6%84%8F%E5%8A%9B--multi-headed-attention
8.2.1 思路https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#821-%E6%80%9D%E8%B7%AF
8.2.2 步骤https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#822-%E6%AD%A5%E9%AA%A4
8.3 代码讲解https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#83-%E4%BB%A3%E7%A0%81%E8%AE%B2%E8%A7%A3
8.4 代码流程总结https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#84-%E4%BB%A3%E7%A0%81%E6%B5%81%E7%A8%8B%E6%80%BB%E7%BB%93
8.5 对比总结https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#85-%E5%AF%B9%E6%AF%94%E6%80%BB%E7%BB%93
九、Transformerhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#%E4%B9%9Dtransformer
9.1 介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#91-%E4%BB%8B%E7%BB%8D
9.2 模型实现https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#92-%E6%A8%A1%E5%9E%8B%E5%AE%9E%E7%8E%B0
9.3 思路分析https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#93-%E6%80%9D%E8%B7%AF%E5%88%86%E6%9E%90
十、入口函数 BertModel()https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#%E5%8D%81%E5%85%A5%E5%8F%A3%E5%87%BD%E6%95%B0-bertmodel
10.1 模型实现https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#101-%E6%A8%A1%E5%9E%8B%E5%AE%9E%E7%8E%B0
10.2 流程介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#102-%E6%B5%81%E7%A8%8B%E4%BB%8B%E7%BB%8D
十一、总结https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode1_modeling.md#%E5%8D%81%E4%B8%80%E6%80%BB%E7%BB%93
【关于 Bert 源码解析II 之 预训练篇】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md
一、动机https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#%E4%B8%80%E5%8A%A8%E6%9C%BA
二、本文框架https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#%E4%BA%8C%E6%9C%AC%E6%96%87%E6%A1%86%E6%9E%B6
三、前言https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#%E4%B8%89%E5%89%8D%E8%A8%80
四、原始语料 预处理模块 （tokenization.py）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#%E5%9B%9B%E5%8E%9F%E5%A7%8B%E8%AF%AD%E6%96%99-%E9%A2%84%E5%A4%84%E7%90%86%E6%A8%A1%E5%9D%97-tokenizationpy
4.1 动机https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#41-%E5%8A%A8%E6%9C%BA
4.2 类别https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#42-%E7%B1%BB%E5%88%AB
4.3 BasicTokenizerhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#43-basictokenizer
4.4 WordpieceTokenizerhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#44-wordpiecetokenizer
4.5 FullTokenizerhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#45-fulltokenizer
五、训练数据生成（create_pretraining_data.py）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#%E4%BA%94%E8%AE%AD%E7%BB%83%E6%95%B0%E6%8D%AE%E7%94%9F%E6%88%90create_pretraining_datapy
5.1 作用https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#51-%E4%BD%9C%E7%94%A8
5.2 参数设置https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#52-%E5%8F%82%E6%95%B0%E8%AE%BE%E7%BD%AE
5.3 main 入口https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#53-main-%E5%85%A5%E5%8F%A3
5.4 定义训练样本类 (TrainingInstance)https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#54-%E5%AE%9A%E4%B9%89%E8%AE%AD%E7%BB%83%E6%A0%B7%E6%9C%AC%E7%B1%BB-traininginstance
5.5 构建训练实例 (create_training_instances)https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#55-%E6%9E%84%E5%BB%BA%E8%AE%AD%E7%BB%83%E5%AE%9E%E4%BE%8B-create_training_instances
5.6 从 document 中抽取 实例（create_instances_from_document）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#56-%E4%BB%8E-document-%E4%B8%AD%E6%8A%BD%E5%8F%96-%E5%AE%9E%E4%BE%8Bcreate_instances_from_document
5.6.1 作用https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#561-%E4%BD%9C%E7%94%A8
5.6.2 代码讲解https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#562-%E4%BB%A3%E7%A0%81%E8%AE%B2%E8%A7%A3
5.6.3 流程https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#563-%E6%B5%81%E7%A8%8B
5.7 随机MASK（create_masked_lm_predictions）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#57-%E9%9A%8F%E6%9C%BAmaskcreate_masked_lm_predictions
5.7.1 介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#571-%E4%BB%8B%E7%BB%8D
5.7.2 代码解析https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#572-%E4%BB%A3%E7%A0%81%E8%A7%A3%E6%9E%90
5.8 保存instance（write_instance_to_example_files）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#58-%E4%BF%9D%E5%AD%98instancewrite_instance_to_example_files
六、预训练https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#%E5%85%AD%E9%A2%84%E8%AE%AD%E7%BB%83
6.1 Masked LM 训练 (get_masked_lm_output)https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#61-masked-lm-%E8%AE%AD%E7%BB%83-get_masked_lm_output
6.2 获取 next sentence prediction（下一句预测） 部分的 loss 以及 log probs (get_next_sentence_output)https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#62-%E8%8E%B7%E5%8F%96-next-sentence-prediction%E4%B8%8B%E4%B8%80%E5%8F%A5%E9%A2%84%E6%B5%8B-%E9%83%A8%E5%88%86%E7%9A%84-loss-%E4%BB%A5%E5%8F%8A-log-probs-get_next_sentence_output
七、测试https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#%E4%B8%83%E6%B5%8B%E8%AF%95
八、总结https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode2_pretraining.md#%E5%85%AB%E6%80%BB%E7%BB%93
【关于 Bert 源码解析III 之 微调篇】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md
一、动机https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#%E4%B8%80%E5%8A%A8%E6%9C%BA
二、本文框架https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#%E4%BA%8C%E6%9C%AC%E6%96%87%E6%A1%86%E6%9E%B6
三、前言https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#%E4%B8%89%E5%89%8D%E8%A8%80
四、参数解析https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#%E5%9B%9B%E5%8F%82%E6%95%B0%E8%A7%A3%E6%9E%90
五、输入数据实例https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#%E4%BA%94%E8%BE%93%E5%85%A5%E6%95%B0%E6%8D%AE%E5%AE%9E%E4%BE%8B
六、特定任务数据处理https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#%E5%85%AD%E7%89%B9%E5%AE%9A%E4%BB%BB%E5%8A%A1%E6%95%B0%E6%8D%AE%E5%A4%84%E7%90%86
6.1 数据处理 接口https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#61-%E6%95%B0%E6%8D%AE%E5%A4%84%E7%90%86-%E6%8E%A5%E5%8F%A3
6.2 推理任务 数据集处理https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#62-%E6%8E%A8%E7%90%86%E4%BB%BB%E5%8A%A1-%E6%95%B0%E6%8D%AE%E9%9B%86%E5%A4%84%E7%90%86
6.3 二分类任务 数据集处理https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#63-%E4%BA%8C%E5%88%86%E7%B1%BB%E4%BB%BB%E5%8A%A1-%E6%95%B0%E6%8D%AE%E9%9B%86%E5%A4%84%E7%90%86
七、examples转换成features （file_based_convert_examples_to_features）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#%E4%B8%83examples%E8%BD%AC%E6%8D%A2%E6%88%90features-file_based_convert_examples_to_features
7.1 单例转化https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#71-%E5%8D%95%E4%BE%8B%E8%BD%AC%E5%8C%96
7.2 单例转化https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#72-%E5%8D%95%E4%BE%8B%E8%BD%AC%E5%8C%96
八、创建模型https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#%E5%85%AB%E5%88%9B%E5%BB%BA%E6%A8%A1%E5%9E%8B
8.1 create_model 创建 分类模型https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#81-create_model-%E5%88%9B%E5%BB%BA-%E5%88%86%E7%B1%BB%E6%A8%A1%E5%9E%8B
8.2 model_fn_builderhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#82-model_fn_builder
九、主入口https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#%E4%B9%9D%E4%B8%BB%E5%85%A5%E5%8F%A3
十、总结https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode3_fineTune.md#%E5%8D%81%E6%80%BB%E7%BB%93
【关于 Bert 源码解析IV 之 句向量生成篇】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode4_word2embedding.md
一、动机https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode4_word2embedding.md#%E4%B8%80%E5%8A%A8%E6%9C%BA
二、本文框架https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode4_word2embedding.md#%E4%BA%8C%E6%9C%AC%E6%96%87%E6%A1%86%E6%9E%B6
三、前言https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode4_word2embedding.md#%E4%B8%89%E5%89%8D%E8%A8%80
四、配置类 （Config）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode4_word2embedding.md#%E5%9B%9B%E9%85%8D%E7%BD%AE%E7%B1%BB-config
五、特征实例类 （InputExample）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode4_word2embedding.md#%E4%BA%94%E7%89%B9%E5%BE%81%E5%AE%9E%E4%BE%8B%E7%B1%BB-inputexample
六、Bert 句向量 类 （BertVector）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode4_word2embedding.md#%E5%85%ADbert-%E5%8F%A5%E5%90%91%E9%87%8F-%E7%B1%BB-bertvector
七、Bert 句向量 生成 实例https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode4_word2embedding.md#%E4%B8%83bert-%E5%8F%A5%E5%90%91%E9%87%8F-%E7%94%9F%E6%88%90-%E5%AE%9E%E4%BE%8B
八、总结https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode4_word2embedding.md#%E5%85%AB%E6%80%BB%E7%BB%93
【关于 Bert 源码解析V 之 文本相似度篇】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md
一、动机https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#%E4%B8%80%E5%8A%A8%E6%9C%BA
二、本文框架https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#%E4%BA%8C%E6%9C%AC%E6%96%87%E6%A1%86%E6%9E%B6
三、前言https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#%E4%B8%89%E5%89%8D%E8%A8%80
四、配置类 （Config）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#%E5%9B%9B%E9%85%8D%E7%BD%AE%E7%B1%BB-config
五、特征实例类 （InputExample）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#%E4%BA%94%E7%89%B9%E5%BE%81%E5%AE%9E%E4%BE%8B%E7%B1%BB-inputexample
六、数据预处理类https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#%E5%85%AD%E6%95%B0%E6%8D%AE%E9%A2%84%E5%A4%84%E7%90%86%E7%B1%BB
6.1 DataProcessorhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#61-dataprocessor
6.2 文本相似度任务 文本预处理 （SimProcessor）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#62-%E6%96%87%E6%9C%AC%E7%9B%B8%E4%BC%BC%E5%BA%A6%E4%BB%BB%E5%8A%A1-%E6%96%87%E6%9C%AC%E9%A2%84%E5%A4%84%E7%90%86-simprocessor
6.2.1 数据格式https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#621-%E6%95%B0%E6%8D%AE%E6%A0%BC%E5%BC%8F
6.2.2 数据预处理类https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#622-%E6%95%B0%E6%8D%AE%E9%A2%84%E5%A4%84%E7%90%86%E7%B1%BB
七、基于 Bert 的 文本相似度 模型https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#%E4%B8%83%E5%9F%BA%E4%BA%8E-bert-%E7%9A%84-%E6%96%87%E6%9C%AC%E7%9B%B8%E4%BC%BC%E5%BA%A6-%E6%A8%A1%E5%9E%8B
八、Bert 相似度 模型 使用https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#%E5%85%ABbert-%E7%9B%B8%E4%BC%BC%E5%BA%A6-%E6%A8%A1%E5%9E%8B-%E4%BD%BF%E7%94%A8
九、总结https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert/bertCode5_similarity.md#%E4%B9%9D%E6%80%BB%E7%BB%93
【关于 小 Bert 模型系列算法】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip
【关于 小 Bert 模型系列算法】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip
一、Bert 模型压缩 动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#%E4%B8%80bert-%E6%A8%A1%E5%9E%8B%E5%8E%8B%E7%BC%A9-%E5%8A%A8%E6%9C%BA%E7%AF%87
二、Bert 模型压缩对比表https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#%E4%BA%8Cbert-%E6%A8%A1%E5%9E%8B%E5%8E%8B%E7%BC%A9%E5%AF%B9%E6%AF%94%E8%A1%A8
三、 Bert 模型压缩方法介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#%E4%B8%89-bert-%E6%A8%A1%E5%9E%8B%E5%8E%8B%E7%BC%A9%E6%96%B9%E6%B3%95%E4%BB%8B%E7%BB%8D
3.1 Bert 模型压缩方法 之 低秩因式分解&跨层参数共享https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#31-bert-%E6%A8%A1%E5%9E%8B%E5%8E%8B%E7%BC%A9%E6%96%B9%E6%B3%95-%E4%B9%8B-%E4%BD%8E%E7%A7%A9%E5%9B%A0%E5%BC%8F%E5%88%86%E8%A7%A3%E8%B7%A8%E5%B1%82%E5%8F%82%E6%95%B0%E5%85%B1%E4%BA%AB
3.1.1 什么是低秩因式分解？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#311-%E4%BB%80%E4%B9%88%E6%98%AF%E4%BD%8E%E7%A7%A9%E5%9B%A0%E5%BC%8F%E5%88%86%E8%A7%A3
3.1.2 什么是跨层参数共享？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#312-%E4%BB%80%E4%B9%88%E6%98%AF%E8%B7%A8%E5%B1%82%E5%8F%82%E6%95%B0%E5%85%B1%E4%BA%AB
3.1.3 ALBERT 所所用的方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#313-albert-%E6%89%80%E6%89%80%E7%94%A8%E7%9A%84%E6%96%B9%E6%B3%95
3.2 Bert 模型压缩方法 之 蒸馏https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#32-bert-%E6%A8%A1%E5%9E%8B%E5%8E%8B%E7%BC%A9%E6%96%B9%E6%B3%95-%E4%B9%8B-%E8%92%B8%E9%A6%8F
3.2.1 什么是蒸馏？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#321-%E4%BB%80%E4%B9%88%E6%98%AF%E8%92%B8%E9%A6%8F
3.2.2 使用 模型蒸馏 的论文？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#322-%E4%BD%BF%E7%94%A8-%E6%A8%A1%E5%9E%8B%E8%92%B8%E9%A6%8F-%E7%9A%84%E8%AE%BA%E6%96%87
3.2.2.1 Extreme Language Model Compression withOptimal Subwords and Shared Projections 【蒸馏】https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#3221-extreme-language-model-compression-withoptimal-subwords-and-shared-projections-%E8%92%B8%E9%A6%8F
3.2.2.2 DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter 【蒸馏】https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#3222-distilbert-a-distilled-version-of-bert-smaller-faster-cheaper-and-lighter-%E8%92%B8%E9%A6%8F
3.2.2.3 FastBERT: a Self-distilling BERT with Adaptive Inference Time 【蒸馏】https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#3223-fastbert-a-self-distilling-bert-with-adaptive-inference-time-%E8%92%B8%E9%A6%8F
3.2.2.4 TinyBERT: Distilling BERT for Natural Language Understanding 【蒸馏】https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#3224-tinybert-distilling-bert-for-natural-language-understanding-%E8%92%B8%E9%A6%8F
3.3 Bert 模型压缩方法 之 量化https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#33-bert-%E6%A8%A1%E5%9E%8B%E5%8E%8B%E7%BC%A9%E6%96%B9%E6%B3%95-%E4%B9%8B-%E9%87%8F%E5%8C%96
3.3.1 什么是量化？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#331-%E4%BB%80%E4%B9%88%E6%98%AF%E9%87%8F%E5%8C%96
3.3.2 Q-BERT: Hessian Based Ultra Low Precision Quantization of BERT 【量化】https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#332--q-bert-hessian-based-ultra-low-precision-quantization-of-bert-%E9%87%8F%E5%8C%96
3.4 Bert 模型压缩方法 之 剪枝https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#34-bert-%E6%A8%A1%E5%9E%8B%E5%8E%8B%E7%BC%A9%E6%96%B9%E6%B3%95-%E4%B9%8B-%E5%89%AA%E6%9E%9D
3.4.1 什么是剪枝？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#341-%E4%BB%80%E4%B9%88%E6%98%AF%E5%89%AA%E6%9E%9D
四、模型压缩存在问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/readme.md#%E5%9B%9B%E6%A8%A1%E5%9E%8B%E5%8E%8B%E7%BC%A9%E5%AD%98%E5%9C%A8%E9%97%AE%E9%A2%98
【关于 Distilling Task-Specific Knowledge from BERT into Simple Neural Networks】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks
一、动机https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA
二、论文思路https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#%E4%BA%8C%E8%AE%BA%E6%96%87%E6%80%9D%E8%B7%AF
三、模型框架讲解【以单句分类任务为例】https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#%E4%B8%89%E6%A8%A1%E5%9E%8B%E6%A1%86%E6%9E%B6%E8%AE%B2%E8%A7%A3%E4%BB%A5%E5%8D%95%E5%8F%A5%E5%88%86%E7%B1%BB%E4%BB%BB%E5%8A%A1%E4%B8%BA%E4%BE%8B
3.1 Teacher 模型（Bert） 微调https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#31-teacher-%E6%A8%A1%E5%9E%8Bbert-%E5%BE%AE%E8%B0%83
3.2 Student 模型（TextCNN、TextRNN）构建https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#32-student-%E6%A8%A1%E5%9E%8Btextcnntextrnn%E6%9E%84%E5%BB%BA
3.2.1 TextRNN 模型构建https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#321-textrnn-%E6%A8%A1%E5%9E%8B%E6%9E%84%E5%BB%BA
3.2.2 TextCNN 模型构建https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#322-textcnn-%E6%A8%A1%E5%9E%8B%E6%9E%84%E5%BB%BA
3.3 Distillation Objectivehttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#33--distillation-objective
四、Data Augmentation for Distillationhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#%E5%9B%9Bdata-augmentation-for-distillation
五、单句分类任务 实验结果分析https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#%E4%BA%94%E5%8D%95%E5%8F%A5%E5%88%86%E7%B1%BB%E4%BB%BB%E5%8A%A1-%E5%AE%9E%E9%AA%8C%E7%BB%93%E6%9E%9C%E5%88%86%E6%9E%90
5.1 数据集介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#51-%E6%95%B0%E6%8D%AE%E9%9B%86%E4%BB%8B%E7%BB%8D
5.2 实验结果分析https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#52-%E5%AE%9E%E9%AA%8C%E7%BB%93%E6%9E%9C%E5%88%86%E6%9E%90
六、总结https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_zip/BERTintoSimpleNeuralNetworks/readme.md#%E5%85%AD%E6%80%BB%E7%BB%93
【关于 大 Bert 模型系列算法】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_big
一、引言https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_big/BERTintoSimpleNeuralNetworks/readme.md#%E4%B8%80%E5%BC%95%E8%A8%80
二、Bert 变大篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_big/BERTintoSimpleNeuralNetworks/readme.md#%E4%BA%8Cbert-%E5%8F%98%E5%A4%A7%E7%AF%87
2.1 认识 XLNet 么？能不能讲一下？ 和 Bert 的 区别在哪里？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_big/BERTintoSimpleNeuralNetworks/readme.md#21-%E8%AE%A4%E8%AF%86-xlnet-%E4%B9%88%E8%83%BD%E4%B8%8D%E8%83%BD%E8%AE%B2%E4%B8%80%E4%B8%8B-%E5%92%8C-bert-%E7%9A%84-%E5%8C%BA%E5%88%AB%E5%9C%A8%E5%93%AA%E9%87%8C
2.2 认识 RoBERTa 么？能不能讲一下？ 和 Bert 的 区别在哪里？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_big/BERTintoSimpleNeuralNetworks/readme.md#22-%E8%AE%A4%E8%AF%86-roberta-%E4%B9%88%E8%83%BD%E4%B8%8D%E8%83%BD%E8%AE%B2%E4%B8%80%E4%B8%8B-%E5%92%8C-bert-%E7%9A%84-%E5%8C%BA%E5%88%AB%E5%9C%A8%E5%93%AA%E9%87%8C
2.3 认识 SpanBERT 么？能不能讲一下？ 和 Bert 的 区别在哪里？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_big/BERTintoSimpleNeuralNetworks/readme.md#23-%E8%AE%A4%E8%AF%86-spanbert-%E4%B9%88%E8%83%BD%E4%B8%8D%E8%83%BD%E8%AE%B2%E4%B8%80%E4%B8%8B-%E5%92%8C-bert-%E7%9A%84-%E5%8C%BA%E5%88%AB%E5%9C%A8%E5%93%AA%E9%87%8C
2.4 认识 MASS 么？能不能讲一下？ 和 Bert 的 区别在哪里？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/PreTraining/bert_big/BERTintoSimpleNeuralNetworks/readme.md#24-%E8%AE%A4%E8%AF%86-mass-%E4%B9%88%E8%83%BD%E4%B8%8D%E8%83%BD%E8%AE%B2%E4%B8%80%E4%B8%8B-%E5%92%8C-bert-%E7%9A%84-%E5%8C%BA%E5%88%AB%E5%9C%A8%E5%93%AA%E9%87%8C
【关于 文本分类】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier
https://github.com/turned123/NLP-Interview-Notes#43-关于-文本分类那些你不知道的事
【关于 文本分类】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification
一、 抽象命题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#%E4%B8%80-%E6%8A%BD%E8%B1%A1%E5%91%BD%E9%A2%98
1.1 分类任务有哪些类别？它们都有什么特征？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#11-%E5%88%86%E7%B1%BB%E4%BB%BB%E5%8A%A1%E6%9C%89%E5%93%AA%E4%BA%9B%E7%B1%BB%E5%88%AB%E5%AE%83%E4%BB%AC%E9%83%BD%E6%9C%89%E4%BB%80%E4%B9%88%E7%89%B9%E5%BE%81
1.2 文本分类任务相较于其他领域的分类任务有何不同之处？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#12-%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E4%BB%BB%E5%8A%A1%E7%9B%B8%E8%BE%83%E4%BA%8E%E5%85%B6%E4%BB%96%E9%A2%86%E5%9F%9F%E7%9A%84%E5%88%86%E7%B1%BB%E4%BB%BB%E5%8A%A1%E6%9C%89%E4%BD%95%E4%B8%8D%E5%90%8C%E4%B9%8B%E5%A4%84
1.3 文本分类任务和文本领域的其他任务相比有何不同之处？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#13-%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E4%BB%BB%E5%8A%A1%E5%92%8C%E6%96%87%E6%9C%AC%E9%A2%86%E5%9F%9F%E7%9A%84%E5%85%B6%E4%BB%96%E4%BB%BB%E5%8A%A1%E7%9B%B8%E6%AF%94%E6%9C%89%E4%BD%95%E4%B8%8D%E5%90%8C%E4%B9%8B%E5%A4%84
1.4 文本分类的过程？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#14-%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E7%9A%84%E8%BF%87%E7%A8%8B
二、数据预处理https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#%E4%BA%8C%E6%95%B0%E6%8D%AE%E9%A2%84%E5%A4%84%E7%90%86
2.1 文本分类任务的数据预处理方法有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#21-%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E4%BB%BB%E5%8A%A1%E7%9A%84%E6%95%B0%E6%8D%AE%E9%A2%84%E5%A4%84%E7%90%86%E6%96%B9%E6%B3%95%E6%9C%89%E5%93%AA%E4%BA%9B
2.2 你使用过哪些分词方法和工具？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#22-%E4%BD%A0%E4%BD%BF%E7%94%A8%E8%BF%87%E5%93%AA%E4%BA%9B%E5%88%86%E8%AF%8D%E6%96%B9%E6%B3%95%E5%92%8C%E5%B7%A5%E5%85%B7
2.3 中文文本分词的方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#23-%E4%B8%AD%E6%96%87%E6%96%87%E6%9C%AC%E5%88%86%E8%AF%8D%E7%9A%84%E6%96%B9%E6%B3%95
2.4 基于字符串匹配的分词方法的原理 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#24-%E5%9F%BA%E4%BA%8E%E5%AD%97%E7%AC%A6%E4%B8%B2%E5%8C%B9%E9%85%8D%E7%9A%84%E5%88%86%E8%AF%8D%E6%96%B9%E6%B3%95%E7%9A%84%E5%8E%9F%E7%90%86-%E6%98%AF%E4%BB%80%E4%B9%88
2.5 统计语言模型如何应用于分词？N-gram最大概率分词？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#25-%E7%BB%9F%E8%AE%A1%E8%AF%AD%E8%A8%80%E6%A8%A1%E5%9E%8B%E5%A6%82%E4%BD%95%E5%BA%94%E7%94%A8%E4%BA%8E%E5%88%86%E8%AF%8Dn-gram%E6%9C%80%E5%A4%A7%E6%A6%82%E7%8E%87%E5%88%86%E8%AF%8D
2.6 基于序列标注的分词方法 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#26-%E5%9F%BA%E4%BA%8E%E5%BA%8F%E5%88%97%E6%A0%87%E6%B3%A8%E7%9A%84%E5%88%86%E8%AF%8D%E6%96%B9%E6%B3%95-%E6%98%AF%E4%BB%80%E4%B9%88
2.7 基于(Bi-)LSTM的词性标注 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#27-%E5%9F%BA%E4%BA%8Ebi-lstm%E7%9A%84%E8%AF%8D%E6%80%A7%E6%A0%87%E6%B3%A8-%E6%98%AF%E4%BB%80%E4%B9%88
2.8 词干提取和词形还原有什么区别？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#28-%E8%AF%8D%E5%B9%B2%E6%8F%90%E5%8F%96%E5%92%8C%E8%AF%8D%E5%BD%A2%E8%BF%98%E5%8E%9F%E6%9C%89%E4%BB%80%E4%B9%88%E5%8C%BA%E5%88%AB
三、特征提取https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#%E4%B8%89%E7%89%B9%E5%BE%81%E6%8F%90%E5%8F%96
3.1 （一个具体的）文本分类任务可以使用哪些特征？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#31-%E4%B8%80%E4%B8%AA%E5%85%B7%E4%BD%93%E7%9A%84%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E4%BB%BB%E5%8A%A1%E5%8F%AF%E4%BB%A5%E4%BD%BF%E7%94%A8%E5%93%AA%E4%BA%9B%E7%89%B9%E5%BE%81
3.2 （对于西文文本）使用单词和使用字母作为特征相比，差异如何？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#32-%E5%AF%B9%E4%BA%8E%E8%A5%BF%E6%96%87%E6%96%87%E6%9C%AC%E4%BD%BF%E7%94%A8%E5%8D%95%E8%AF%8D%E5%92%8C%E4%BD%BF%E7%94%A8%E5%AD%97%E6%AF%8D%E4%BD%9C%E4%B8%BA%E7%89%B9%E5%BE%81%E7%9B%B8%E6%AF%94%E5%B7%AE%E5%BC%82%E5%A6%82%E4%BD%95
3.3 能不能简单介绍下词袋模型？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#33-%E8%83%BD%E4%B8%8D%E8%83%BD%E7%AE%80%E5%8D%95%E4%BB%8B%E7%BB%8D%E4%B8%8B%E8%AF%8D%E8%A2%8B%E6%A8%A1%E5%9E%8B
3.4 n-gram 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#34-n-gram-%E7%AF%87
3.4.1 什么是n元语法？为什么要用n-gram？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#341-%E4%BB%80%E4%B9%88%E6%98%AFn%E5%85%83%E8%AF%AD%E6%B3%95%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E7%94%A8n-gram
3.4.2 n-gram算法的局限性是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#342-n-gram%E7%AE%97%E6%B3%95%E7%9A%84%E5%B1%80%E9%99%90%E6%80%A7%E6%98%AF%E4%BB%80%E4%B9%88
3.5 主题建模篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#35-%E4%B8%BB%E9%A2%98%E5%BB%BA%E6%A8%A1%E7%AF%87
3.5.1 介绍一下主题建模任务？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#351-%E4%BB%8B%E7%BB%8D%E4%B8%80%E4%B8%8B%E4%B8%BB%E9%A2%98%E5%BB%BA%E6%A8%A1%E4%BB%BB%E5%8A%A1
3.5.2 主题建模的常用方法https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#352-%E4%B8%BB%E9%A2%98%E5%BB%BA%E6%A8%A1%E7%9A%84%E5%B8%B8%E7%94%A8%E6%96%B9%E6%B3%95
3.5.3 TF-IDF算法是做什么的？简单介绍下TF-IDF算法https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#353-tf-idf%E7%AE%97%E6%B3%95%E6%98%AF%E5%81%9A%E4%BB%80%E4%B9%88%E7%9A%84%E7%AE%80%E5%8D%95%E4%BB%8B%E7%BB%8D%E4%B8%8Btf-idf%E7%AE%97%E6%B3%95
3.5.4 tf-idf高意味着什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#354-tf-idf%E9%AB%98%E6%84%8F%E5%91%B3%E7%9D%80%E4%BB%80%E4%B9%88
3.5.5 tf-idf的不足之处https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#355-tf-idf%E7%9A%84%E4%B8%8D%E8%B6%B3%E4%B9%8B%E5%A4%84
3.6 文本相似度篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#36-%E6%96%87%E6%9C%AC%E7%9B%B8%E4%BC%BC%E5%BA%A6%E7%AF%87
3.6.1 如何计算两段文本之间的距离？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#361-%E5%A6%82%E4%BD%95%E8%AE%A1%E7%AE%97%E4%B8%A4%E6%AE%B5%E6%96%87%E6%9C%AC%E4%B9%8B%E9%97%B4%E7%9A%84%E8%B7%9D%E7%A6%BB
3.6.2 什么是jaccard距离？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#362-%E4%BB%80%E4%B9%88%E6%98%AFjaccard%E8%B7%9D%E7%A6%BB
3.6.3 Dice系数和Jaccard系数的区别？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#363-dice%E7%B3%BB%E6%95%B0%E5%92%8Cjaccard%E7%B3%BB%E6%95%B0%E7%9A%84%E5%8C%BA%E5%88%AB
3.6.4 同样是编辑距离，莱文斯坦距离和汉明距离的区别在哪里？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#364-%E5%90%8C%E6%A0%B7%E6%98%AF%E7%BC%96%E8%BE%91%E8%B7%9D%E7%A6%BB%E8%8E%B1%E6%96%87%E6%96%AF%E5%9D%A6%E8%B7%9D%E7%A6%BB%E5%92%8C%E6%B1%89%E6%98%8E%E8%B7%9D%E7%A6%BB%E7%9A%84%E5%8C%BA%E5%88%AB%E5%9C%A8%E5%93%AA%E9%87%8C
3.6.5 写一下计算编辑距离（莱温斯坦距离）的编程题吧？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#365-%E5%86%99%E4%B8%80%E4%B8%8B%E8%AE%A1%E7%AE%97%E7%BC%96%E8%BE%91%E8%B7%9D%E7%A6%BB%E8%8E%B1%E6%B8%A9%E6%96%AF%E5%9D%A6%E8%B7%9D%E7%A6%BB%E7%9A%84%E7%BC%96%E7%A8%8B%E9%A2%98%E5%90%A7
四、模型篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#%E5%9B%9B%E6%A8%A1%E5%9E%8B%E7%AF%87
4.1 fastText 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#41-fasttext-%E7%AF%87
4.1.1 fastText的分类过程？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#411-fasttext%E7%9A%84%E5%88%86%E7%B1%BB%E8%BF%87%E7%A8%8B
4.1.2 fastText的优点？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#412-fasttext%E7%9A%84%E4%BC%98%E7%82%B9
4.2 TextCNN 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#42-textcnn-%E7%AF%87
4.2.1 TextCNN进行文本分类的过程?https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#421-textcnn%E8%BF%9B%E8%A1%8C%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E7%9A%84%E8%BF%87%E7%A8%8B
4.2.2 TextCNN可以调整哪些参数？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#422-textcnn%E5%8F%AF%E4%BB%A5%E8%B0%83%E6%95%B4%E5%93%AA%E4%BA%9B%E5%8F%82%E6%95%B0
4.2.3 使用CNN作为文本分类器时，不同通道channels对应着文本的什么信息？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#423-%E4%BD%BF%E7%94%A8cnn%E4%BD%9C%E4%B8%BA%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E5%99%A8%E6%97%B6%E4%B8%8D%E5%90%8C%E9%80%9A%E9%81%93channels%E5%AF%B9%E5%BA%94%E7%9D%80%E6%96%87%E6%9C%AC%E7%9A%84%E4%BB%80%E4%B9%88%E4%BF%A1%E6%81%AF
4.2.4 TextCNN中卷积核的长与宽代表了什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#424-textcnn%E4%B8%AD%E5%8D%B7%E7%A7%AF%E6%A0%B8%E7%9A%84%E9%95%BF%E4%B8%8E%E5%AE%BD%E4%BB%A3%E8%A1%A8%E4%BA%86%E4%BB%80%E4%B9%88
4.2.5 在TextCNN中的pooling操作与一般CNN的pooling操作有何不同？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#425-%E5%9C%A8textcnn%E4%B8%AD%E7%9A%84pooling%E6%93%8D%E4%BD%9C%E4%B8%8E%E4%B8%80%E8%88%ACcnn%E7%9A%84pooling%E6%93%8D%E4%BD%9C%E6%9C%89%E4%BD%95%E4%B8%8D%E5%90%8C
4.2.6 TextCNN的局限性？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#426-textcnn%E7%9A%84%E5%B1%80%E9%99%90%E6%80%A7
4.3 DPCNN 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#43-dpcnn-%E7%AF%87
4.3.1 如何解决长文本分类任务？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#431-%E5%A6%82%E4%BD%95%E8%A7%A3%E5%86%B3%E9%95%BF%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E4%BB%BB%E5%8A%A1
4.3.2 简单介绍DPCNN模型相较于TextCNN的改进？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#432-%E7%AE%80%E5%8D%95%E4%BB%8B%E7%BB%8Ddpcnn%E6%A8%A1%E5%9E%8B%E7%9B%B8%E8%BE%83%E4%BA%8Etextcnn%E7%9A%84%E6%94%B9%E8%BF%9B
4.4 TextRCNN 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#44-textrcnn-%E7%AF%87
4.4.1 简要介绍TextRCNN相较于TextCNN的改进？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#441-%E7%AE%80%E8%A6%81%E4%BB%8B%E7%BB%8Dtextrcnn%E7%9B%B8%E8%BE%83%E4%BA%8Etextcnn%E7%9A%84%E6%94%B9%E8%BF%9B
4.5 RNN+Attention 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#45-rnnattention-%E7%AF%87
4.5.1 RNN+Attention进行文本分类任务的思路，以及为什么要加Attention / 注意力机制如何应用于文本分类领域？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#451-rnnattention%E8%BF%9B%E8%A1%8C%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E4%BB%BB%E5%8A%A1%E7%9A%84%E6%80%9D%E8%B7%AF%E4%BB%A5%E5%8F%8A%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E5%8A%A0attention--%E6%B3%A8%E6%84%8F%E5%8A%9B%E6%9C%BA%E5%88%B6%E5%A6%82%E4%BD%95%E5%BA%94%E7%94%A8%E4%BA%8E%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E9%A2%86%E5%9F%9F
4.6 GNN 图神经网络篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#46-gnn-%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C%E7%AF%87
4.6.1 GNN 图神经网络如何应用于文本分类领域？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#461-gnn-%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C%E5%A6%82%E4%BD%95%E5%BA%94%E7%94%A8%E4%BA%8E%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E9%A2%86%E5%9F%9F
4.7 Transformer 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#47-transformer-%E7%AF%87
4.7.1 基于Transformer的预训练模型如何应用于文本分类领域？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#471-%E5%9F%BA%E4%BA%8Etransformer%E7%9A%84%E9%A2%84%E8%AE%AD%E7%BB%83%E6%A8%A1%E5%9E%8B%E5%A6%82%E4%BD%95%E5%BA%94%E7%94%A8%E4%BA%8E%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E9%A2%86%E5%9F%9F
4.8 预训练模型 篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#48-%E9%A2%84%E8%AE%AD%E7%BB%83%E6%A8%A1%E5%9E%8B-%E7%AF%87
4.8.1 你了解哪些预训练模型？它们的特点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#481-%E4%BD%A0%E4%BA%86%E8%A7%A3%E5%93%AA%E4%BA%9B%E9%A2%84%E8%AE%AD%E7%BB%83%E6%A8%A1%E5%9E%8B%E5%AE%83%E4%BB%AC%E7%9A%84%E7%89%B9%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
五、损失函数https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#%E4%BA%94%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0
5.1 激活函数sigmoid篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#51-%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0sigmoid%E7%AF%87
5.1.1 二分类问题使用的激活函数sigmoid简介？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#511-%E4%BA%8C%E5%88%86%E7%B1%BB%E9%97%AE%E9%A2%98%E4%BD%BF%E7%94%A8%E7%9A%84%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0sigmoid%E7%AE%80%E4%BB%8B
5.1.2 Sigmod的缺点是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#512-sigmod%E7%9A%84%E7%BC%BA%E7%82%B9%E6%98%AF%E4%BB%80%E4%B9%88
5.2 激活函数softmax篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#52-%E6%BF%80%E6%B4%BB%E5%87%BD%E6%95%B0softmax%E7%AF%87
5.2.1 softmax函数是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#521-softmax%E5%87%BD%E6%95%B0%E6%98%AF%E4%BB%80%E4%B9%88
5.2.2 softmax函数怎么求导？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#522-softmax%E5%87%BD%E6%95%B0%E6%80%8E%E4%B9%88%E6%B1%82%E5%AF%BC
5.3 分类问题使用的损失函数还有有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#53-%E5%88%86%E7%B1%BB%E9%97%AE%E9%A2%98%E4%BD%BF%E7%94%A8%E7%9A%84%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0%E8%BF%98%E6%9C%89%E6%9C%89%E5%93%AA%E4%BA%9B
六、模型评估和算法比较https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#%E5%85%AD%E6%A8%A1%E5%9E%8B%E8%AF%84%E4%BC%B0%E5%92%8C%E7%AE%97%E6%B3%95%E6%AF%94%E8%BE%83
6.1 文本分类任务使用的评估算法和指标有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#61-%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB%E4%BB%BB%E5%8A%A1%E4%BD%BF%E7%94%A8%E7%9A%84%E8%AF%84%E4%BC%B0%E7%AE%97%E6%B3%95%E5%92%8C%E6%8C%87%E6%A0%87%E6%9C%89%E5%93%AA%E4%BA%9B
6.2 简单介绍混淆矩阵和kappa？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/TextClassification/readme.md#62-%E7%AE%80%E5%8D%95%E4%BB%8B%E7%BB%8D%E6%B7%B7%E6%B7%86%E7%9F%A9%E9%98%B5%E5%92%8Ckappa
【关于 文本分类 trick 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick
一、数据预处理问题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#%E4%B8%80%E6%95%B0%E6%8D%AE%E9%A2%84%E5%A4%84%E7%90%86%E9%97%AE%E9%A2%98
1.1 vocab 构建问题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#11-vocab-%E6%9E%84%E5%BB%BA%E9%97%AE%E9%A2%98
1.2 模型输入问题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#12-%E6%A8%A1%E5%9E%8B%E8%BE%93%E5%85%A5%E9%97%AE%E9%A2%98
1.3 噪声数据处理问题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#13-%E5%99%AA%E5%A3%B0%E6%95%B0%E6%8D%AE%E5%A4%84%E7%90%86%E9%97%AE%E9%A2%98
1.4 中文任务分词问题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#14-%E4%B8%AD%E6%96%87%E4%BB%BB%E5%8A%A1%E5%88%86%E8%AF%8D%E9%97%AE%E9%A2%98
1.5 停用词处理问题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#15-%E5%81%9C%E7%94%A8%E8%AF%8D%E5%A4%84%E7%90%86%E9%97%AE%E9%A2%98
二、模型篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#%E4%BA%8C%E6%A8%A1%E5%9E%8B%E7%AF%87
2.1 模型选择https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#21-%E6%A8%A1%E5%9E%8B%E9%80%89%E6%8B%A9
2.2 词向量选择https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#22-%E8%AF%8D%E5%90%91%E9%87%8F%E9%80%89%E6%8B%A9
2.3 字 or 词向量 预训练https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#23-%E5%AD%97-or-%E8%AF%8D%E5%90%91%E9%87%8F-%E9%A2%84%E8%AE%AD%E7%BB%83
三、参数篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#%E4%B8%89%E5%8F%82%E6%95%B0%E7%AF%87
3.1 正则化https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#31-%E6%AD%A3%E5%88%99%E5%8C%96
3.2 学习率https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#32-%E5%AD%A6%E4%B9%A0%E7%8E%87
四、任务篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#%E5%9B%9B%E4%BB%BB%E5%8A%A1%E7%AF%87
4.1 二分类问题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#41-%E4%BA%8C%E5%88%86%E7%B1%BB%E9%97%AE%E9%A2%98
4.2 多标签分类https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#42-%E5%A4%9A%E6%A0%87%E7%AD%BE%E5%88%86%E7%B1%BB
4.3 长文本问题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#43-%E9%95%BF%E6%96%87%E6%9C%AC%E9%97%AE%E9%A2%98
4.4 鲁棒性问题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#44-%E9%B2%81%E6%A3%92%E6%80%A7%E9%97%AE%E9%A2%98
五、标签体系构建https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#%E4%BA%94%E6%A0%87%E7%AD%BE%E4%BD%93%E7%B3%BB%E6%9E%84%E5%BB%BA
5.1 标签体系构建https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#51-%E6%A0%87%E7%AD%BE%E4%BD%93%E7%B3%BB%E6%9E%84%E5%BB%BA
5.2 标签体系合理性评估https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#52-%E6%A0%87%E7%AD%BE%E4%BD%93%E7%B3%BB%E5%90%88%E7%90%86%E6%80%A7%E8%AF%84%E4%BC%B0
六、策略构建https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#%E5%85%AD%E7%AD%96%E7%95%A5%E6%9E%84%E5%BB%BA
6.1 算法策略构建https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#61-%E7%AE%97%E6%B3%95%E7%AD%96%E7%95%A5%E6%9E%84%E5%BB%BA
6.2 特征挖掘策略https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#62-%E7%89%B9%E5%BE%81%E6%8C%96%E6%8E%98%E7%AD%96%E7%95%A5
6.3 数据不均衡问题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#63-%E6%95%B0%E6%8D%AE%E4%B8%8D%E5%9D%87%E8%A1%A1%E9%97%AE%E9%A2%98
6.3.1 重采样（re-sampling）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#631-%E9%87%8D%E9%87%87%E6%A0%B7re-sampling
6.3.2 重加权（re-weighting）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#632-%E9%87%8D%E5%8A%A0%E6%9D%83re-weighting
6.3.3 数据增强https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#633-%E6%95%B0%E6%8D%AE%E5%A2%9E%E5%BC%BA
6.4 预训练模型融合角度https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#64-%E9%A2%84%E8%AE%AD%E7%BB%83%E6%A8%A1%E5%9E%8B%E8%9E%8D%E5%90%88%E8%A7%92%E5%BA%A6
6.5 灾难性遗忘问题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#65-%E7%81%BE%E9%9A%BE%E6%80%A7%E9%81%97%E5%BF%98%E9%97%AE%E9%A2%98
6.6 小模型大智慧https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#66-%E5%B0%8F%E6%A8%A1%E5%9E%8B%E5%A4%A7%E6%99%BA%E6%85%A7
6.6.1 模型蒸馏https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#661-%E6%A8%A1%E5%9E%8B%E8%92%B8%E9%A6%8F
6.6.2 数据蒸馏https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/textclassifier/ClassifierTrick/readme.md#662-%E6%95%B0%E6%8D%AE%E8%92%B8%E9%A6%8F
【关于 文本匹配】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch
https://github.com/turned123/NLP-Interview-Notes#44-关于-文本匹配那些你不知道的事
【关于 文本匹配模型 ESIM 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/ESIM
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/ESIM/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
二、ESIM 模型篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/ESIM/readme.md#%E4%BA%8Cesim-%E6%A8%A1%E5%9E%8B%E7%AF%87
2.1 模型介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/ESIM/readme.md#21-%E6%A8%A1%E5%9E%8B%E4%BB%8B%E7%BB%8D
2.2 Input Encodinghttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/ESIM/readme.md#22-input-encoding
2.3 Local Inference Modelinghttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/ESIM/readme.md#23-local-inference-modeling
2.4 Inference Compositionhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/ESIM/readme.md#24-inference-composition
2.5 Predictionhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/ESIM/readme.md#25-prediction
2.6 模型训练https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/ESIM/readme.md#26-%E6%A8%A1%E5%9E%8B%E8%AE%AD%E7%BB%83
【关于 语义相似度匹配任务中的 BERT】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/bert_similairity
一、Sentence Pair Classification Task：使用 [CLS]https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/bert_similairity/readme.md#%E4%B8%80sentence-pair-classification-task%E4%BD%BF%E7%94%A8-cls
二、cosine similairityhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/bert_similairity/readme.md#%E4%BA%8Ccosine-similairity
三、长短文本的区别https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/bert_similairity/readme.md#%E4%B8%89%E9%95%BF%E7%9F%AD%E6%96%87%E6%9C%AC%E7%9A%84%E5%8C%BA%E5%88%AB
四、sentence/word embeddinghttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/bert_similairity/readme.md#%E5%9B%9Bsentenceword-embedding
五、siamese network 方式https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMatch/bert_similairity/readme.md#%E4%BA%94siamese-network-%E6%96%B9%E5%BC%8F
【关于 问答系统】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA
https://github.com/turned123/NLP-Interview-Notes#45-关于-问答系统那些你不知道的事
【关于 FAQ 检索式问答系统】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ
https://github.com/turned123/NLP-Interview-Notes#451-关于-faq-检索式问答系统-那些你不知道的事
【关于 FAQ 检索式问答系统】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ
一、动机https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA
1.1 问答系统的动机？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#11-%E9%97%AE%E7%AD%94%E7%B3%BB%E7%BB%9F%E7%9A%84%E5%8A%A8%E6%9C%BA
1.2 问答系统 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#12-%E9%97%AE%E7%AD%94%E7%B3%BB%E7%BB%9F-%E6%98%AF%E4%BB%80%E4%B9%88
二、FAQ 检索式问答系统介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#%E4%BA%8Cfaq-%E6%A3%80%E7%B4%A2%E5%BC%8F%E9%97%AE%E7%AD%94%E7%B3%BB%E7%BB%9F%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1 FAQ 检索式问答系统 是 什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#21-faq-%E6%A3%80%E7%B4%A2%E5%BC%8F%E9%97%AE%E7%AD%94%E7%B3%BB%E7%BB%9F-%E6%98%AF-%E4%BB%80%E4%B9%88
2.2 query 匹配标准 QA 的核心是什么?https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#22-query-%E5%8C%B9%E9%85%8D%E6%A0%87%E5%87%86-qa-%E7%9A%84%E6%A0%B8%E5%BF%83%E6%98%AF%E4%BB%80%E4%B9%88
三、FAQ 检索式问答系统 方案篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#%E4%B8%89faq-%E6%A3%80%E7%B4%A2%E5%BC%8F%E9%97%AE%E7%AD%94%E7%B3%BB%E7%BB%9F-%E6%96%B9%E6%A1%88%E7%AF%87
3.1 常用 方案有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#31-%E5%B8%B8%E7%94%A8-%E6%96%B9%E6%A1%88%E6%9C%89%E5%93%AA%E4%BA%9B
3.2 为什么 QQ 匹配比较常用？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#32-%E4%B8%BA%E4%BB%80%E4%B9%88-qq-%E5%8C%B9%E9%85%8D%E6%AF%94%E8%BE%83%E5%B8%B8%E7%94%A8
3.2.1 QQ 匹配的优点有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#321-qq-%E5%8C%B9%E9%85%8D%E7%9A%84%E4%BC%98%E7%82%B9%E6%9C%89%E5%93%AA%E4%BA%9B
3.2.2 QQ 匹配的语义空间是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#322-qq-%E5%8C%B9%E9%85%8D%E7%9A%84%E8%AF%AD%E4%B9%89%E7%A9%BA%E9%97%B4%E6%98%AF%E4%BB%80%E4%B9%88
3.2.3 QQ 匹配的语料的稳定性是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#323-qq-%E5%8C%B9%E9%85%8D%E7%9A%84%E8%AF%AD%E6%96%99%E7%9A%84%E7%A8%B3%E5%AE%9A%E6%80%A7%E6%98%AF%E4%BB%80%E4%B9%88
3.2.4 QQ 匹配的业务回答与算法模型的解耦是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#324-qq-%E5%8C%B9%E9%85%8D%E7%9A%84%E4%B8%9A%E5%8A%A1%E5%9B%9E%E7%AD%94%E4%B8%8E%E7%AE%97%E6%B3%95%E6%A8%A1%E5%9E%8B%E7%9A%84%E8%A7%A3%E8%80%A6%E6%98%AF%E4%BB%80%E4%B9%88
3.2.5 QQ 匹配的新问题发现与去重是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#325-qq-%E5%8C%B9%E9%85%8D%E7%9A%84%E6%96%B0%E9%97%AE%E9%A2%98%E5%8F%91%E7%8E%B0%E4%B8%8E%E5%8E%BB%E9%87%8D%E6%98%AF%E4%BB%80%E4%B9%88
3.2.6 QQ 匹配的上线运行速度是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#326-qq-%E5%8C%B9%E9%85%8D%E7%9A%84%E4%B8%8A%E7%BA%BF%E8%BF%90%E8%A1%8C%E9%80%9F%E5%BA%A6%E6%98%AF%E4%BB%80%E4%B9%88
3.3 QQ 匹配一般处理流程是怎么样？ 【假设 标准问题库 已处理好】https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#33--qq-%E5%8C%B9%E9%85%8D%E4%B8%80%E8%88%AC%E5%A4%84%E7%90%86%E6%B5%81%E7%A8%8B%E6%98%AF%E6%80%8E%E4%B9%88%E6%A0%B7-%E5%81%87%E8%AE%BE-%E6%A0%87%E5%87%86%E9%97%AE%E9%A2%98%E5%BA%93-%E5%B7%B2%E5%A4%84%E7%90%86%E5%A5%BD
四、FAQ 标准问题库构建篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#%E5%9B%9Bfaq-%E6%A0%87%E5%87%86%E9%97%AE%E9%A2%98%E5%BA%93%E6%9E%84%E5%BB%BA%E7%AF%87
4.1 如何发现 FAQ 中标准问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#41-%E5%A6%82%E4%BD%95%E5%8F%91%E7%8E%B0-faq-%E4%B8%AD%E6%A0%87%E5%87%86%E9%97%AE%E9%A2%98
4.2 FAQ 如何做拆分？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#42-faq-%E5%A6%82%E4%BD%95%E5%81%9A%E6%8B%86%E5%88%86
4.3 FAQ 如何做合并？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#43-faq-%E5%A6%82%E4%BD%95%E5%81%9A%E5%90%88%E5%B9%B6
4.4 FAQ 标准库如何实时更新？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#44-faq-%E6%A0%87%E5%87%86%E5%BA%93%E5%A6%82%E4%BD%95%E5%AE%9E%E6%97%B6%E6%9B%B4%E6%96%B0
五、FAQ 标准问题库答案优化篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#%E4%BA%94faq-%E6%A0%87%E5%87%86%E9%97%AE%E9%A2%98%E5%BA%93%E7%AD%94%E6%A1%88%E4%BC%98%E5%8C%96%E7%AF%87
5.1 FAQ 标准问题库答案如何优化？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#51-faq-%E6%A0%87%E5%87%86%E9%97%AE%E9%A2%98%E5%BA%93%E7%AD%94%E6%A1%88%E5%A6%82%E4%BD%95%E4%BC%98%E5%8C%96
参考https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/FAQ/readme.md#%E5%8F%82%E8%80%83
【关于 问答系统工具篇】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss
https://github.com/turned123/NLP-Interview-Notes#452-关于-问答系统工具篇-那些你不知道的事
【关于 Faiss 】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 传统的相似度算法所存在的问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#11-%E4%BC%A0%E7%BB%9F%E7%9A%84%E7%9B%B8%E4%BC%BC%E5%BA%A6%E7%AE%97%E6%B3%95%E6%89%80%E5%AD%98%E5%9C%A8%E7%9A%84%E9%97%AE%E9%A2%98
二、介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#%E4%BA%8C%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1 什么是 Faiss ？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#21-%E4%BB%80%E4%B9%88%E6%98%AF-faiss-
2.2 Faiss 如何使用？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#22-faiss-%E5%A6%82%E4%BD%95%E4%BD%BF%E7%94%A8
2.3 Faiss原理与核心算法https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#23-faiss%E5%8E%9F%E7%90%86%E4%B8%8E%E6%A0%B8%E5%BF%83%E7%AE%97%E6%B3%95
三、Faiss 实战篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#%E4%B8%89faiss-%E5%AE%9E%E6%88%98%E7%AF%87
3.1 Faiss 如何安装？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#31-faiss-%E5%A6%82%E4%BD%95%E5%AE%89%E8%A3%85
3.2 Faiss 的索引Index有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#32-faiss-%E7%9A%84%E7%B4%A2%E5%BC%95index%E6%9C%89%E5%93%AA%E4%BA%9B
3.3 Faiss 的索引Index都怎么用？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#33-faiss-%E7%9A%84%E7%B4%A2%E5%BC%95index%E9%83%BD%E6%80%8E%E4%B9%88%E7%94%A8
3.3.1 数据预备https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#331-%E6%95%B0%E6%8D%AE%E9%A2%84%E5%A4%87
3.3.2 暴力美学 IndexFlatL2https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#332-%E6%9A%B4%E5%8A%9B%E7%BE%8E%E5%AD%A6-indexflatl2
3.3.3 闪电侠 IndexIVFFlathttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#333-%E9%97%AA%E7%94%B5%E4%BE%A0-indexivfflat
3.3.4 内存管家 IndexIVFPQhttps://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#334-%E5%86%85%E5%AD%98%E7%AE%A1%E5%AE%B6-indexivfpq
3.4 Faiss 然后使用 GPU？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#34-faiss-%E7%84%B6%E5%90%8E%E4%BD%BF%E7%94%A8-gpu
四、 Faiss 对比篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#%E5%9B%9B-faiss-%E5%AF%B9%E6%AF%94%E7%AF%87
4.1 sklearn cosine_similarity 和 Faiss 哪家强https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/QA/Faiss/readme.md#41-sklearn-cosine_similarity--%E5%92%8C-faiss--%E5%93%AA%E5%AE%B6%E5%BC%BA
【关于 对话系统】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem
https://github.com/turned123/NLP-Interview-Notes#46-关于-对话系统那些你不知道的事
【关于 对话系统】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem
一、对话系统 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#%E4%B8%80%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F-%E4%BB%8B%E7%BB%8D%E7%AF%87
1.1 对话系统有哪几种？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#11-%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F%E6%9C%89%E5%93%AA%E5%87%A0%E7%A7%8D
1.2 这几种对话系统的区别？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#12-%E8%BF%99%E5%87%A0%E7%A7%8D%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F%E7%9A%84%E5%8C%BA%E5%88%AB
二、多轮对话系统 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#%E4%BA%8C%E5%A4%9A%E8%BD%AE%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F-%E4%BB%8B%E7%BB%8D%E7%AF%87
2.1 为什么要用 多轮对话系统？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#21-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E7%94%A8-%E5%A4%9A%E8%BD%AE%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F
2.2 常见的多轮对话系统解决方案是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#22-%E5%B8%B8%E8%A7%81%E7%9A%84%E5%A4%9A%E8%BD%AE%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F%E8%A7%A3%E5%86%B3%E6%96%B9%E6%A1%88%E6%98%AF%E4%BB%80%E4%B9%88
三、任务型对话系统 介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#%E4%B8%89%E4%BB%BB%E5%8A%A1%E5%9E%8B%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F-%E4%BB%8B%E7%BB%8D%E7%AF%87
3.1 什么是任务型对话系统？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#31-%E4%BB%80%E4%B9%88%E6%98%AF%E4%BB%BB%E5%8A%A1%E5%9E%8B%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F
3.2 任务型对话系统的流程是怎么样？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#32-%E4%BB%BB%E5%8A%A1%E5%9E%8B%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F%E7%9A%84%E6%B5%81%E7%A8%8B%E6%98%AF%E6%80%8E%E4%B9%88%E6%A0%B7
3.3 任务型对话系统 语言理解（SLU）篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#33-%E4%BB%BB%E5%8A%A1%E5%9E%8B%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F-%E8%AF%AD%E8%A8%80%E7%90%86%E8%A7%A3slu%E7%AF%87
3.3.1 什么是 语言理解（SLU）？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#331-%E4%BB%80%E4%B9%88%E6%98%AF-%E8%AF%AD%E8%A8%80%E7%90%86%E8%A7%A3slu
3.3.2 语言理解（SLU）的输入输出是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#332-%E8%AF%AD%E8%A8%80%E7%90%86%E8%A7%A3slu%E7%9A%84%E8%BE%93%E5%85%A5%E8%BE%93%E5%87%BA%E6%98%AF%E4%BB%80%E4%B9%88
3.3.3 语言理解（SLU）所使用的技术是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#333-%E8%AF%AD%E8%A8%80%E7%90%86%E8%A7%A3slu%E6%89%80%E4%BD%BF%E7%94%A8%E7%9A%84%E6%8A%80%E6%9C%AF%E6%98%AF%E4%BB%80%E4%B9%88
3.4 任务型对话系统 DST（对话状态跟踪）篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#34-%E4%BB%BB%E5%8A%A1%E5%9E%8B%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F-dst%E5%AF%B9%E8%AF%9D%E7%8A%B6%E6%80%81%E8%B7%9F%E8%B8%AA%E7%AF%87
3.4.1 什么是 DST（对话状态跟踪）？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#341-%E4%BB%80%E4%B9%88%E6%98%AF-dst%E5%AF%B9%E8%AF%9D%E7%8A%B6%E6%80%81%E8%B7%9F%E8%B8%AA
3.4.2 DST（对话状态跟踪）的输入输出是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#342-dst%E5%AF%B9%E8%AF%9D%E7%8A%B6%E6%80%81%E8%B7%9F%E8%B8%AA%E7%9A%84%E8%BE%93%E5%85%A5%E8%BE%93%E5%87%BA%E6%98%AF%E4%BB%80%E4%B9%88
3.4.3 DST（对话状态跟踪）存在问题和解决方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#343-dst%E5%AF%B9%E8%AF%9D%E7%8A%B6%E6%80%81%E8%B7%9F%E8%B8%AA%E5%AD%98%E5%9C%A8%E9%97%AE%E9%A2%98%E5%92%8C%E8%A7%A3%E5%86%B3%E6%96%B9%E6%B3%95
3.4.4 DST（对话状态跟踪）实现方式是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#344-dst%E5%AF%B9%E8%AF%9D%E7%8A%B6%E6%80%81%E8%B7%9F%E8%B8%AA%E5%AE%9E%E7%8E%B0%E6%96%B9%E5%BC%8F%E6%98%AF%E4%BB%80%E4%B9%88
3.5 任务型对话系统 DPO（对话策略学习）篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#35-%E4%BB%BB%E5%8A%A1%E5%9E%8B%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F-dpo%E5%AF%B9%E8%AF%9D%E7%AD%96%E7%95%A5%E5%AD%A6%E4%B9%A0%E7%AF%87
3.5.1 DPO（对话策略学习）是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#351-dpo%E5%AF%B9%E8%AF%9D%E7%AD%96%E7%95%A5%E5%AD%A6%E4%B9%A0%E6%98%AF%E4%BB%80%E4%B9%88
3.5.2 DPO（对话策略学习）的输入输出是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#352-dpo%E5%AF%B9%E8%AF%9D%E7%AD%96%E7%95%A5%E5%AD%A6%E4%B9%A0%E7%9A%84%E8%BE%93%E5%85%A5%E8%BE%93%E5%87%BA%E6%98%AF%E4%BB%80%E4%B9%88
3.5.3 DPO（对话策略学习）的实现方法是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#353-dpo%E5%AF%B9%E8%AF%9D%E7%AD%96%E7%95%A5%E5%AD%A6%E4%B9%A0%E7%9A%84%E5%AE%9E%E7%8E%B0%E6%96%B9%E6%B3%95%E6%98%AF%E4%BB%80%E4%B9%88
3.6 任务型对话系统 NLG（自然语言生成）篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#36-%E4%BB%BB%E5%8A%A1%E5%9E%8B%E5%AF%B9%E8%AF%9D%E7%B3%BB%E7%BB%9F-nlg%E8%87%AA%E7%84%B6%E8%AF%AD%E8%A8%80%E7%94%9F%E6%88%90%E7%AF%87
3.6.1 NLG（自然语言生成）是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#361-nlg%E8%87%AA%E7%84%B6%E8%AF%AD%E8%A8%80%E7%94%9F%E6%88%90%E6%98%AF%E4%BB%80%E4%B9%88
3.6.2 NLG（自然语言生成）的输入输出是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#362-nlg%E8%87%AA%E7%84%B6%E8%AF%AD%E8%A8%80%E7%94%9F%E6%88%90%E7%9A%84%E8%BE%93%E5%85%A5%E8%BE%93%E5%87%BA%E6%98%AF%E4%BB%80%E4%B9%88
3.6.3 NLG（自然语言生成）的实现方式？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/readme.md#363-nlg%E8%87%AA%E7%84%B6%E8%AF%AD%E8%A8%80%E7%94%9F%E6%88%90%E7%9A%84%E5%AE%9E%E7%8E%B0%E6%96%B9%E5%BC%8F
【关于 RASA】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/DialogueSystem/Rasa
【关于 知识图谱】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG
https://github.com/turned123/NLP-Interview-Notes#47-关于-知识图谱那些你不知道的事
【关于 知识图谱】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG
https://github.com/turned123/NLP-Interview-Notes#471-关于-知识图谱-那些你不知道的事
【关于 知识图谱】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG
一、知识图谱简介https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#%E4%B8%80%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1%E7%AE%80%E4%BB%8B
1.1 引言https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#11-%E5%BC%95%E8%A8%80
1.2 什么是知识图谱呢？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#12-%E4%BB%80%E4%B9%88%E6%98%AF%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1%E5%91%A2
1.2.1 什么是图（Graph）呢？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#121-%E4%BB%80%E4%B9%88%E6%98%AF%E5%9B%BEgraph%E5%91%A2
1.2.2 什么是 Schema 呢？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#122-%E4%BB%80%E4%B9%88%E6%98%AF-schema-%E5%91%A2
1.3 知识图谱的类别有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#13-%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1%E7%9A%84%E7%B1%BB%E5%88%AB%E6%9C%89%E5%93%AA%E4%BA%9B
1.4 知识图谱的价值在哪呢？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#14-%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1%E7%9A%84%E4%BB%B7%E5%80%BC%E5%9C%A8%E5%93%AA%E5%91%A2
二、怎么构建知识图谱呢？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#%E4%BA%8C%E6%80%8E%E4%B9%88%E6%9E%84%E5%BB%BA%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1%E5%91%A2
2.1 知识图谱的数据来源于哪里？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#21-%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1%E7%9A%84%E6%95%B0%E6%8D%AE%E6%9D%A5%E6%BA%90%E4%BA%8E%E5%93%AA%E9%87%8C
2.2 信息抽取的难点在哪里？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#22-%E4%BF%A1%E6%81%AF%E6%8A%BD%E5%8F%96%E7%9A%84%E9%9A%BE%E7%82%B9%E5%9C%A8%E5%93%AA%E9%87%8C
2.3 构建知识图谱所涉及的技术？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#23-%E6%9E%84%E5%BB%BA%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1%E6%89%80%E6%B6%89%E5%8F%8A%E7%9A%84%E6%8A%80%E6%9C%AF
2.4、知识图谱的具体构建技术是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#24%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1%E7%9A%84%E5%85%B7%E4%BD%93%E6%9E%84%E5%BB%BA%E6%8A%80%E6%9C%AF%E6%98%AF%E4%BB%80%E4%B9%88
2.4.1 实体命名识别（Named Entity Recognition）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#241-%E5%AE%9E%E4%BD%93%E5%91%BD%E5%90%8D%E8%AF%86%E5%88%ABnamed-entity-recognition
2.4.2 关系抽取（Relation Extraction）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#242-%E5%85%B3%E7%B3%BB%E6%8A%BD%E5%8F%96relation-extraction
2.4.3 实体统一（Entity Resolution）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#243-%E5%AE%9E%E4%BD%93%E7%BB%9F%E4%B8%80entity-resolution
2.4.4 指代消解（Disambiguation）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#244-%E6%8C%87%E4%BB%A3%E6%B6%88%E8%A7%A3disambiguation
三、知识图谱怎么存储？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#%E4%B8%89%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1%E6%80%8E%E4%B9%88%E5%AD%98%E5%82%A8
四、知识图谱可以做什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#%E5%9B%9B%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1%E5%8F%AF%E4%BB%A5%E5%81%9A%E4%BB%80%E4%B9%88
参考资料https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/readme.md#%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99
【关于 KBQA】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA
https://github.com/turned123/NLP-Interview-Notes#472-关于-kbqa那些你不知道的事
【关于 KBQA】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA
一、基于词典和规则的方法https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#%E4%B8%80%E5%9F%BA%E4%BA%8E%E8%AF%8D%E5%85%B8%E5%92%8C%E8%A7%84%E5%88%99%E7%9A%84%E6%96%B9%E6%B3%95
1.1 介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#11-%E4%BB%8B%E7%BB%8D
1.1.1 开源知识图谱https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#111-%E5%BC%80%E6%BA%90%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1
1.1.2 代表项目https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#112-%E4%BB%A3%E8%A1%A8%E9%A1%B9%E7%9B%AE
1.2 流程https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#12-%E6%B5%81%E7%A8%8B
1.2.1. 句子输入https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#121-%E5%8F%A5%E5%AD%90%E8%BE%93%E5%85%A5
1.2.2. 问句解析https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#122-%E9%97%AE%E5%8F%A5%E8%A7%A3%E6%9E%90
1.2.3. 查询语句生成https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#123-%E6%9F%A5%E8%AF%A2%E8%AF%AD%E5%8F%A5%E7%94%9F%E6%88%90
1.2.4. 查询数据库和结果生成https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#124-%E6%9F%A5%E8%AF%A2%E6%95%B0%E6%8D%AE%E5%BA%93%E5%92%8C%E7%BB%93%E6%9E%9C%E7%94%9F%E6%88%90
二、基于信息抽取的方法https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#%E4%BA%8C%E5%9F%BA%E4%BA%8E%E4%BF%A1%E6%81%AF%E6%8A%BD%E5%8F%96%E7%9A%84%E6%96%B9%E6%B3%95
2.1 介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#21-%E4%BB%8B%E7%BB%8D
2.1.1 开源知识图谱介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#211-%E5%BC%80%E6%BA%90%E7%9F%A5%E8%AF%86%E5%9B%BE%E8%B0%B1%E4%BB%8B%E7%BB%8D
2.1.2 评测标准https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#212-%E8%AF%84%E6%B5%8B%E6%A0%87%E5%87%86
2.2 流程https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#22-%E6%B5%81%E7%A8%8B
2.2.1. 分类单跳和多跳问句https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#221-%E5%88%86%E7%B1%BB%E5%8D%95%E8%B7%B3%E5%92%8C%E5%A4%9A%E8%B7%B3%E9%97%AE%E5%8F%A5
2.2.2. 分类链式问句（二分类）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#222-%E5%88%86%E7%B1%BB%E9%93%BE%E5%BC%8F%E9%97%AE%E5%8F%A5%E4%BA%8C%E5%88%86%E7%B1%BB
2.2.3. 主谓宾分类（三分类）https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#223-%E4%B8%BB%E8%B0%93%E5%AE%BE%E5%88%86%E7%B1%BB%E4%B8%89%E5%88%86%E7%B1%BB
2.2.4. 实体提及（mention）识别https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#224-%E5%AE%9E%E4%BD%93%E6%8F%90%E5%8F%8Amention%E8%AF%86%E5%88%AB
2.2.5. 关系分类 (语义相似度计算，二分类问题)https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#225-%E5%85%B3%E7%B3%BB%E5%88%86%E7%B1%BB-%E8%AF%AD%E4%B9%89%E7%9B%B8%E4%BC%BC%E5%BA%A6%E8%AE%A1%E7%AE%97%E4%BA%8C%E5%88%86%E7%B1%BB%E9%97%AE%E9%A2%98
2.2.6. 实体链指 【实体消歧】https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#226-%E5%AE%9E%E4%BD%93%E9%93%BE%E6%8C%87-%E5%AE%9E%E4%BD%93%E6%B6%88%E6%AD%A7
2.2.7. 候选查询路径生成及文本匹配https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#227-%E5%80%99%E9%80%89%E6%9F%A5%E8%AF%A2%E8%B7%AF%E5%BE%84%E7%94%9F%E6%88%90%E5%8F%8A%E6%96%87%E6%9C%AC%E5%8C%B9%E9%85%8D
2.2.8. 实体桥接及答案检索https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/KBQA/readme.md#228-%E5%AE%9E%E4%BD%93%E6%A1%A5%E6%8E%A5%E5%8F%8A%E7%AD%94%E6%A1%88%E6%A3%80%E7%B4%A2
【关于 Neo4j】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j
https://github.com/turned123/NLP-Interview-Notes#473-关于-neo4j那些你不知道的事
【关于 Neo4j】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j
一、Neo4J 介绍与安装https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#%E4%B8%80neo4j-%E4%BB%8B%E7%BB%8D%E4%B8%8E%E5%AE%89%E8%A3%85
1.1 引言https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#11-%E5%BC%95%E8%A8%80
1.2 Neo4J 怎么下载？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#12-neo4j-%E6%80%8E%E4%B9%88%E4%B8%8B%E8%BD%BD
1.3 Neo4J 怎么安装？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#13-neo4j-%E6%80%8E%E4%B9%88%E5%AE%89%E8%A3%85
1.4 Neo4J Web 界面 介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#14-neo4j-web-%E7%95%8C%E9%9D%A2-%E4%BB%8B%E7%BB%8D
1.5 Cypher查询语言是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#15-cypher%E6%9F%A5%E8%AF%A2%E8%AF%AD%E8%A8%80%E6%98%AF%E4%BB%80%E4%B9%88
二、Neo4J 增删查改篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#%E4%BA%8Cneo4j-%E5%A2%9E%E5%88%A0%E6%9F%A5%E6%94%B9%E7%AF%87
2.1 引言https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#21-%E5%BC%95%E8%A8%80
2.2 Neo4j 怎么创建节点？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#22-neo4j-%E6%80%8E%E4%B9%88%E5%88%9B%E5%BB%BA%E8%8A%82%E7%82%B9
2.3 Neo4j 怎么创建关系？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#23-neo4j-%E6%80%8E%E4%B9%88%E5%88%9B%E5%BB%BA%E5%85%B3%E7%B3%BB
2.4 Neo4j 怎么创建 出生地关系？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#24-neo4j-%E6%80%8E%E4%B9%88%E5%88%9B%E5%BB%BA-%E5%87%BA%E7%94%9F%E5%9C%B0%E5%85%B3%E7%B3%BB
2.5 Neo4j 怎么查询？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#25-neo4j-%E6%80%8E%E4%B9%88%E6%9F%A5%E8%AF%A2
2.6 Neo4j 怎么删除和修改？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#26-neo4j-%E6%80%8E%E4%B9%88%E5%88%A0%E9%99%A4%E5%92%8C%E4%BF%AE%E6%94%B9
三、如何利用 Python 操作 Neo4j 图数据库？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#%E4%B8%89%E5%A6%82%E4%BD%95%E5%88%A9%E7%94%A8-python-%E6%93%8D%E4%BD%9C-neo4j-%E5%9B%BE%E6%95%B0%E6%8D%AE%E5%BA%93
3.1 neo4j模块：执行CQL ( cypher ) 语句是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#31-neo4j%E6%A8%A1%E5%9D%97%E6%89%A7%E8%A1%8Ccql--cypher--%E8%AF%AD%E5%8F%A5%E6%98%AF%E4%BB%80%E4%B9%88
3.2 py2neo模块是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#32-py2neo%E6%A8%A1%E5%9D%97%E6%98%AF%E4%BB%80%E4%B9%88
四、数据导入 Neo4j 图数据库篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#%E5%9B%9B%E6%95%B0%E6%8D%AE%E5%AF%BC%E5%85%A5-neo4j-%E5%9B%BE%E6%95%B0%E6%8D%AE%E5%BA%93%E7%AF%87
参考资料https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KG/neo4j/readme.md#%E5%8F%82%E8%80%83%E8%B5%84%E6%96%99
【关于 文本摘要】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary
https://github.com/turned123/NLP-Interview-Notes#48-关于-文本摘要-那些你不知道的事
【关于 文本摘要】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 什么是文本摘要？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#11-%E4%BB%80%E4%B9%88%E6%98%AF%E6%96%87%E6%9C%AC%E6%91%98%E8%A6%81
1.2 文本摘要技术有哪些类型？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#12-%E6%96%87%E6%9C%AC%E6%91%98%E8%A6%81%E6%8A%80%E6%9C%AF%E6%9C%89%E5%93%AA%E4%BA%9B%E7%B1%BB%E5%9E%8B
二、抽取式摘要篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#%E4%BA%8C%E6%8A%BD%E5%8F%96%E5%BC%8F%E6%91%98%E8%A6%81%E7%AF%87
2.1 抽取式摘要是怎么做的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#21-%E6%8A%BD%E5%8F%96%E5%BC%8F%E6%91%98%E8%A6%81%E6%98%AF%E6%80%8E%E4%B9%88%E5%81%9A%E7%9A%84
2.1.1 句子重要性评估算法有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#211-%E5%8F%A5%E5%AD%90%E9%87%8D%E8%A6%81%E6%80%A7%E8%AF%84%E4%BC%B0%E7%AE%97%E6%B3%95%E6%9C%89%E5%93%AA%E4%BA%9B
2.1.2 基于约束的摘要生成方法有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#212-%E5%9F%BA%E4%BA%8E%E7%BA%A6%E6%9D%9F%E7%9A%84%E6%91%98%E8%A6%81%E7%94%9F%E6%88%90%E6%96%B9%E6%B3%95%E6%9C%89%E5%93%AA%E4%BA%9B
2.1.3 TextTeaser算法是怎么抽取摘要的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#213-textteaser%E7%AE%97%E6%B3%95%E6%98%AF%E6%80%8E%E4%B9%88%E6%8A%BD%E5%8F%96%E6%91%98%E8%A6%81%E7%9A%84
2.1.4 TextRank算法是怎么抽取摘要的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#214-textrank%E7%AE%97%E6%B3%95%E6%98%AF%E6%80%8E%E4%B9%88%E6%8A%BD%E5%8F%96%E6%91%98%E8%A6%81%E7%9A%84
2.2 抽取式摘要的可读性问题是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#22-%E6%8A%BD%E5%8F%96%E5%BC%8F%E6%91%98%E8%A6%81%E7%9A%84%E5%8F%AF%E8%AF%BB%E6%80%A7%E9%97%AE%E9%A2%98%E6%98%AF%E4%BB%80%E4%B9%88
三、压缩式摘要篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#%E4%B8%89%E5%8E%8B%E7%BC%A9%E5%BC%8F%E6%91%98%E8%A6%81%E7%AF%87
3.1 压缩式摘要是怎么做的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#31-%E5%8E%8B%E7%BC%A9%E5%BC%8F%E6%91%98%E8%A6%81%E6%98%AF%E6%80%8E%E4%B9%88%E5%81%9A%E7%9A%84
四、生成式摘要篇https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#%E5%9B%9B%E7%94%9F%E6%88%90%E5%BC%8F%E6%91%98%E8%A6%81%E7%AF%87
4.1 生成式摘要是怎么做的？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#41-%E7%94%9F%E6%88%90%E5%BC%8F%E6%91%98%E8%A6%81%E6%98%AF%E6%80%8E%E4%B9%88%E5%81%9A%E7%9A%84
4.2 生成式摘要存在哪些问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#42-%E7%94%9F%E6%88%90%E5%BC%8F%E6%91%98%E8%A6%81%E5%AD%98%E5%9C%A8%E5%93%AA%E4%BA%9B%E9%97%AE%E9%A2%98
4.3 Pointer-generator network解决了什么问题？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#43-pointer-generator-network%E8%A7%A3%E5%86%B3%E4%BA%86%E4%BB%80%E4%B9%88%E9%97%AE%E9%A2%98
五、摘要质量评估方法https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#%E4%BA%94%E6%91%98%E8%A6%81%E8%B4%A8%E9%87%8F%E8%AF%84%E4%BC%B0%E6%96%B9%E6%B3%95
5.1 摘要质量的评估方法有哪些类型？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#51-%E6%91%98%E8%A6%81%E8%B4%A8%E9%87%8F%E7%9A%84%E8%AF%84%E4%BC%B0%E6%96%B9%E6%B3%95%E6%9C%89%E5%93%AA%E4%BA%9B%E7%B1%BB%E5%9E%8B
5.2 什么是ROUGE？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#52-%E4%BB%80%E4%B9%88%E6%98%AFrouge
5.3 几种ROUGE指标之间的区别是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#53-%E5%87%A0%E7%A7%8Drouge%E6%8C%87%E6%A0%87%E4%B9%8B%E9%97%B4%E7%9A%84%E5%8C%BA%E5%88%AB%E6%98%AF%E4%BB%80%E4%B9%88
5.4 BLEU和ROUGE有什么不同？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/summary/readme.md#54-bleu%E5%92%8Crouge%E6%9C%89%E4%BB%80%E4%B9%88%E4%B8%8D%E5%90%8C
【关于 知识表示学习】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation
https://github.com/turned123/NLP-Interview-Notes#49-关于-知识表示学习那些你不知道的事
【关于 数据挖掘】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation
一. 理论及研究现状https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#%E4%B8%80-%E7%90%86%E8%AE%BA%E5%8F%8A%E7%A0%94%E7%A9%B6%E7%8E%B0%E7%8A%B6
1.1 理论https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#11-%E7%90%86%E8%AE%BA
1.1.1 知识表示学习的基本概念https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#111-%E7%9F%A5%E8%AF%86%E8%A1%A8%E7%A4%BA%E5%AD%A6%E4%B9%A0%E7%9A%84%E5%9F%BA%E6%9C%AC%E6%A6%82%E5%BF%B5
1.1.2 知识表示的理论基础https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#112-%E7%9F%A5%E8%AF%86%E8%A1%A8%E7%A4%BA%E7%9A%84%E7%90%86%E8%AE%BA%E5%9F%BA%E7%A1%80
1.1.3 知识表示学习的典型应用https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#113-%E7%9F%A5%E8%AF%86%E8%A1%A8%E7%A4%BA%E5%AD%A6%E4%B9%A0%E7%9A%84%E5%85%B8%E5%9E%8B%E5%BA%94%E7%94%A8
1.1.4 知识表示学习的主要优点https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#114-%E7%9F%A5%E8%AF%86%E8%A1%A8%E7%A4%BA%E5%AD%A6%E4%B9%A0%E7%9A%84%E4%B8%BB%E8%A6%81%E4%BC%98%E7%82%B9
1.2 研究现状https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#12-%E7%A0%94%E7%A9%B6%E7%8E%B0%E7%8A%B6
二. 常见面试题https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#%E4%BA%8C-%E5%B8%B8%E8%A7%81%E9%9D%A2%E8%AF%95%E9%A2%98
2.1 Q： 知识表示相对于one-hot表示的优势是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#21-q-%E7%9F%A5%E8%AF%86%E8%A1%A8%E7%A4%BA%E7%9B%B8%E5%AF%B9%E4%BA%8Eone-hot%E8%A1%A8%E7%A4%BA%E7%9A%84%E4%BC%98%E5%8A%BF%E6%98%AF%E4%BB%80%E4%B9%88
2.2 Q：有哪些文本表示模型?它们各有什么优缺点?https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#22-q%E6%9C%89%E5%93%AA%E4%BA%9B%E6%96%87%E6%9C%AC%E8%A1%A8%E7%A4%BA%E6%A8%A1%E5%9E%8B%E5%AE%83%E4%BB%AC%E5%90%84%E6%9C%89%E4%BB%80%E4%B9%88%E4%BC%98%E7%BC%BA%E7%82%B9
2.3 Q：word2vec与LDA模型之间的区别和联系？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#23-qword2vec%E4%B8%8Elda%E6%A8%A1%E5%9E%8B%E4%B9%8B%E9%97%B4%E7%9A%84%E5%8C%BA%E5%88%AB%E5%92%8C%E8%81%94%E7%B3%BB
2.4 Q：介绍下词向量空间中的平移不变现象？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#24-q%E4%BB%8B%E7%BB%8D%E4%B8%8B%E8%AF%8D%E5%90%91%E9%87%8F%E7%A9%BA%E9%97%B4%E4%B8%AD%E7%9A%84%E5%B9%B3%E7%A7%BB%E4%B8%8D%E5%8F%98%E7%8E%B0%E8%B1%A1
2.5 Q：简要介绍下TransE模型的思想及优点？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#25-q%E7%AE%80%E8%A6%81%E4%BB%8B%E7%BB%8D%E4%B8%8Btranse%E6%A8%A1%E5%9E%8B%E7%9A%84%E6%80%9D%E6%83%B3%E5%8F%8A%E4%BC%98%E7%82%B9
2.6 Q：解释一下为什么TransE模型用于复杂关系建模时的性能较差？https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#26-q%E8%A7%A3%E9%87%8A%E4%B8%80%E4%B8%8B%E4%B8%BA%E4%BB%80%E4%B9%88transe%E6%A8%A1%E5%9E%8B%E7%94%A8%E4%BA%8E%E5%A4%8D%E6%9D%82%E5%85%B3%E7%B3%BB%E5%BB%BA%E6%A8%A1%E6%97%B6%E7%9A%84%E6%80%A7%E8%83%BD%E8%BE%83%E5%B7%AE
2.7 Q：简述TransH、TransR和TransD模型的思想https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#27-q%E7%AE%80%E8%BF%B0transhtransr%E5%92%8Ctransd%E6%A8%A1%E5%9E%8B%E7%9A%84%E6%80%9D%E6%83%B3
2.8 Q：简述deepwalk和node2vec模型的思想及其优点https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#28-q%E7%AE%80%E8%BF%B0deepwalk%E5%92%8Cnode2vec%E6%A8%A1%E5%9E%8B%E7%9A%84%E6%80%9D%E6%83%B3%E5%8F%8A%E5%85%B6%E4%BC%98%E7%82%B9
2.9 Q：简述Line模型的思想https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#29-q%E7%AE%80%E8%BF%B0line%E6%A8%A1%E5%9E%8B%E7%9A%84%E6%80%9D%E6%83%B3
参考文献https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/KnowledgeRepresentation/readme.md#%E5%8F%82%E8%80%83%E6%96%87%E7%8C%AE
【关于 数据挖掘】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMining
https://github.com/turned123/NLP-Interview-Notes#410--关于-数据挖掘那些你不知道的事
【关于 数据挖掘】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/NLPinterview/TextMining
一、什么是文本挖掘？https://github.com/turned123/NLP-Interview-Notes#%E4%B8%80%E4%BB%80%E4%B9%88%E6%98%AF%E6%96%87%E6%9C%AC%E6%8C%96%E6%8E%98
二、文本挖掘的作用是什么？https://github.com/turned123/NLP-Interview-Notes#%E4%BA%8C%E6%96%87%E6%9C%AC%E6%8C%96%E6%8E%98%E7%9A%84%E4%BD%9C%E7%94%A8%E6%98%AF%E4%BB%80%E4%B9%88
三、文本预处理https://github.com/turned123/NLP-Interview-Notes#%E4%B8%89%E6%96%87%E6%9C%AC%E9%A2%84%E5%A4%84%E7%90%86
3.1 中文分词https://github.com/turned123/NLP-Interview-Notes#31-%E4%B8%AD%E6%96%87%E5%88%86%E8%AF%8D
3.2 去停用词https://github.com/turned123/NLP-Interview-Notes#32-%E5%8E%BB%E5%81%9C%E7%94%A8%E8%AF%8D
3.3 低频词和高频词处理https://github.com/turned123/NLP-Interview-Notes#33-%E4%BD%8E%E9%A2%91%E8%AF%8D%E5%92%8C%E9%AB%98%E9%A2%91%E8%AF%8D%E5%A4%84%E7%90%86
3.4 计算 N-gram 【这里采用 Bigrams】https://github.com/turned123/NLP-Interview-Notes#34-%E8%AE%A1%E7%AE%97-n-gram-%E8%BF%99%E9%87%8C%E9%87%87%E7%94%A8-bigrams
四、文本挖掘https://github.com/turned123/NLP-Interview-Notes#%E5%9B%9B%E6%96%87%E6%9C%AC%E6%8C%96%E6%8E%98
4.1 关键词提取https://github.com/turned123/NLP-Interview-Notes#41-%E5%85%B3%E9%94%AE%E8%AF%8D%E6%8F%90%E5%8F%96
4.2 LDA 主题模型分析https://github.com/turned123/NLP-Interview-Notes#42-lda-%E4%B8%BB%E9%A2%98%E6%A8%A1%E5%9E%8B%E5%88%86%E6%9E%90
4.3 情绪分析&LDA主题模型交叉分析https://github.com/turned123/NLP-Interview-Notes#43-%E6%83%85%E7%BB%AA%E5%88%86%E6%9E%90lda%E4%B8%BB%E9%A2%98%E6%A8%A1%E5%9E%8B%E4%BA%A4%E5%8F%89%E5%88%86%E6%9E%90
4.4 ATM 模型https://github.com/turned123/NLP-Interview-Notes#44-atm-%E6%A8%A1%E5%9E%8B
4.5 词向量训练及关联词分析https://github.com/turned123/NLP-Interview-Notes#45-%E8%AF%8D%E5%90%91%E9%87%8F%E8%AE%AD%E7%BB%83%E5%8F%8A%E5%85%B3%E8%81%94%E8%AF%8D%E5%88%86%E6%9E%90
4.6 词聚类与词分类https://github.com/turned123/NLP-Interview-Notes#46-%E8%AF%8D%E8%81%9A%E7%B1%BB%E4%B8%8E%E8%AF%8D%E5%88%86%E7%B1%BB
4.7 文本分类https://github.com/turned123/NLP-Interview-Notes#47-%E6%96%87%E6%9C%AC%E5%88%86%E7%B1%BB
4.8 文本聚类https://github.com/turned123/NLP-Interview-Notes#48-%E6%96%87%E6%9C%AC%E8%81%9A%E7%B1%BB
4.9 信息检索https://github.com/turned123/NLP-Interview-Notes#49-%E4%BF%A1%E6%81%AF%E6%A3%80%E7%B4%A2
【关于 NLP 技巧】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick
https://github.com/turned123/NLP-Interview-Notes#五关于-nlp-技巧那些你不知道的事
【关于 少样本问题】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem
https://github.com/turned123/NLP-Interview-Notes#51-关于-少样本问题那些你不知道的事
【关于 EDA 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 什么是 数据增强？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#11-%E4%BB%80%E4%B9%88%E6%98%AF-%E6%95%B0%E6%8D%AE%E5%A2%9E%E5%BC%BA
1.2 为什么需要 数据增强？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#12-%E4%B8%BA%E4%BB%80%E4%B9%88%E9%9C%80%E8%A6%81-%E6%95%B0%E6%8D%AE%E5%A2%9E%E5%BC%BA
二、常见的数据增强方法篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#%E4%BA%8C%E5%B8%B8%E8%A7%81%E7%9A%84%E6%95%B0%E6%8D%AE%E5%A2%9E%E5%BC%BA%E6%96%B9%E6%B3%95%E7%AF%87
2.1 词汇替换篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#21-%E8%AF%8D%E6%B1%87%E6%9B%BF%E6%8D%A2%E7%AF%87
2.1.1 什么是基于词典的替换方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#211-%E4%BB%80%E4%B9%88%E6%98%AF%E5%9F%BA%E4%BA%8E%E8%AF%8D%E5%85%B8%E7%9A%84%E6%9B%BF%E6%8D%A2%E6%96%B9%E6%B3%95
2.1.2 什么是基于词向量的替换方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#212-%E4%BB%80%E4%B9%88%E6%98%AF%E5%9F%BA%E4%BA%8E%E8%AF%8D%E5%90%91%E9%87%8F%E7%9A%84%E6%9B%BF%E6%8D%A2%E6%96%B9%E6%B3%95
2.1.3 什么是基于 MLM 的替换方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#213-%E4%BB%80%E4%B9%88%E6%98%AF%E5%9F%BA%E4%BA%8E-mlm-%E7%9A%84%E6%9B%BF%E6%8D%A2%E6%96%B9%E6%B3%95
2.1.4 什么是基于 TF-IDF 的词替换？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#214-%E4%BB%80%E4%B9%88%E6%98%AF%E5%9F%BA%E4%BA%8E-tf-idf-%E7%9A%84%E8%AF%8D%E6%9B%BF%E6%8D%A2
2.2 词汇插入篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#22-%E8%AF%8D%E6%B1%87%E6%8F%92%E5%85%A5%E7%AF%87
2.2.1 什么是随机插入法？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#221-%E4%BB%80%E4%B9%88%E6%98%AF%E9%9A%8F%E6%9C%BA%E6%8F%92%E5%85%A5%E6%B3%95
2.3 词汇交换篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#23-%E8%AF%8D%E6%B1%87%E4%BA%A4%E6%8D%A2%E7%AF%87
2.3.1 什么是随机交换法？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#231-%E4%BB%80%E4%B9%88%E6%98%AF%E9%9A%8F%E6%9C%BA%E4%BA%A4%E6%8D%A2%E6%B3%95
2.4 词汇删除篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#24-%E8%AF%8D%E6%B1%87%E5%88%A0%E9%99%A4%E7%AF%87
2.4.1 什么是随机删除法？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#241-%E4%BB%80%E4%B9%88%E6%98%AF%E9%9A%8F%E6%9C%BA%E5%88%A0%E9%99%A4%E6%B3%95
2.5 回译篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#25-%E5%9B%9E%E8%AF%91%E7%AF%87
2.5.1 什么是回译法？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#251-%E4%BB%80%E4%B9%88%E6%98%AF%E5%9B%9E%E8%AF%91%E6%B3%95
2.6 交叉增强篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#26-%E4%BA%A4%E5%8F%89%E5%A2%9E%E5%BC%BA%E7%AF%87
2.6.1 什么是 交叉增强篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#261-%E4%BB%80%E4%B9%88%E6%98%AF-%E4%BA%A4%E5%8F%89%E5%A2%9E%E5%BC%BA%E7%AF%87
2.7 语法树篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#27-%E8%AF%AD%E6%B3%95%E6%A0%91%E7%AF%87
2.7.1 什么是语法树操作？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#271-%E4%BB%80%E4%B9%88%E6%98%AF%E8%AF%AD%E6%B3%95%E6%A0%91%E6%93%8D%E4%BD%9C
2.8 对抗增强篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#28-%E5%AF%B9%E6%8A%97%E5%A2%9E%E5%BC%BA%E7%AF%87
2.8.1 什么是对抗增强？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/EDA/eda.md#281-%E4%BB%80%E4%B9%88%E6%98%AF%E5%AF%B9%E6%8A%97%E5%A2%9E%E5%BC%BA
【关于 主动学习 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md
一、动机篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA%E7%AF%87
1.1 主动学习是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md#11-%E4%B8%BB%E5%8A%A8%E5%AD%A6%E4%B9%A0%E6%98%AF%E4%BB%80%E4%B9%88
1.2 为什么需要主动学习？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md#12-%E4%B8%BA%E4%BB%80%E4%B9%88%E9%9C%80%E8%A6%81%E4%B8%BB%E5%8A%A8%E5%AD%A6%E4%B9%A0
二、主动学习篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md#%E4%BA%8C%E4%B8%BB%E5%8A%A8%E5%AD%A6%E4%B9%A0%E7%AF%87
2.1 主动学习的思路是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md#21-%E4%B8%BB%E5%8A%A8%E5%AD%A6%E4%B9%A0%E7%9A%84%E6%80%9D%E8%B7%AF%E6%98%AF%E4%BB%80%E4%B9%88
2.2 主动学习方法 的价值点在哪里？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md#22-%E4%B8%BB%E5%8A%A8%E5%AD%A6%E4%B9%A0%E6%96%B9%E6%B3%95-%E7%9A%84%E4%BB%B7%E5%80%BC%E7%82%B9%E5%9C%A8%E5%93%AA%E9%87%8C
三、样本选取策略篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md#%E4%B8%89%E6%A0%B7%E6%9C%AC%E9%80%89%E5%8F%96%E7%AD%96%E7%95%A5%E7%AF%87
3.1 以未标记样本的获取方式的差别进行划分https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md#31-%E4%BB%A5%E6%9C%AA%E6%A0%87%E8%AE%B0%E6%A0%B7%E6%9C%AC%E7%9A%84%E8%8E%B7%E5%8F%96%E6%96%B9%E5%BC%8F%E7%9A%84%E5%B7%AE%E5%88%AB%E8%BF%9B%E8%A1%8C%E5%88%92%E5%88%86
3.2 测试集内选取“信息”量最大的数据标记https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md#32-%E6%B5%8B%E8%AF%95%E9%9B%86%E5%86%85%E9%80%89%E5%8F%96%E4%BF%A1%E6%81%AF%E9%87%8F%E6%9C%80%E5%A4%A7%E7%9A%84%E6%95%B0%E6%8D%AE%E6%A0%87%E8%AE%B0
3.2.1 测试集内选取“信息”量最大的数据标记https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md#321-%E6%B5%8B%E8%AF%95%E9%9B%86%E5%86%85%E9%80%89%E5%8F%96%E4%BF%A1%E6%81%AF%E9%87%8F%E6%9C%80%E5%A4%A7%E7%9A%84%E6%95%B0%E6%8D%AE%E6%A0%87%E8%AE%B0
3.2.2 依赖不确定度的样本选取策略（Uncertainty Sampling, US）https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md#322-%E4%BE%9D%E8%B5%96%E4%B8%8D%E7%A1%AE%E5%AE%9A%E5%BA%A6%E7%9A%84%E6%A0%B7%E6%9C%AC%E9%80%89%E5%8F%96%E7%AD%96%E7%95%A5uncertainty-sampling-us
3.2.3 基于委员会查询的方法（Query-By-Committee，QBC）https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/activeLearn/readme.md#323-%E5%9F%BA%E4%BA%8E%E5%A7%94%E5%91%98%E4%BC%9A%E6%9F%A5%E8%AF%A2%E7%9A%84%E6%96%B9%E6%B3%95query-by-committeeqbc
【关于 数据增强 之 对抗训练】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md
一、介绍篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md#%E4%B8%80%E4%BB%8B%E7%BB%8D%E7%AF%87
1.1 什么是 对抗训练 ？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md#11-%E4%BB%80%E4%B9%88%E6%98%AF-%E5%AF%B9%E6%8A%97%E8%AE%AD%E7%BB%83-
1.2 为什么 对抗训练 能够 提高模型效果？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md#12-%E4%B8%BA%E4%BB%80%E4%B9%88-%E5%AF%B9%E6%8A%97%E8%AE%AD%E7%BB%83-%E8%83%BD%E5%A4%9F-%E6%8F%90%E9%AB%98%E6%A8%A1%E5%9E%8B%E6%95%88%E6%9E%9C
1.3 对抗训练 有什么特点？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md#13--%E5%AF%B9%E6%8A%97%E8%AE%AD%E7%BB%83-%E6%9C%89%E4%BB%80%E4%B9%88%E7%89%B9%E7%82%B9
1.4 对抗训练 的作用?https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md#14-%E5%AF%B9%E6%8A%97%E8%AE%AD%E7%BB%83-%E7%9A%84%E4%BD%9C%E7%94%A8
二、概念篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md#%E4%BA%8C%E6%A6%82%E5%BF%B5%E7%AF%87
2.1 对抗训练的基本概念?https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md#21-%E5%AF%B9%E6%8A%97%E8%AE%AD%E7%BB%83%E7%9A%84%E5%9F%BA%E6%9C%AC%E6%A6%82%E5%BF%B5
2.2 如何计算扰动?https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md#22-%E5%A6%82%E4%BD%95%E8%AE%A1%E7%AE%97%E6%89%B0%E5%8A%A8
2.3 如何优化?https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md#23-%E5%A6%82%E4%BD%95%E4%BC%98%E5%8C%96
三、实战篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md#%E4%B8%89%E5%AE%9E%E6%88%98%E7%AF%87
3.1 NLP 中经典对抗训练 之 Fast Gradient Method（FGM）https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md#31-nlp-%E4%B8%AD%E7%BB%8F%E5%85%B8%E5%AF%B9%E6%8A%97%E8%AE%AD%E7%BB%83-%E4%B9%8B--fast-gradient-methodfgm
3.2 NLP 中经典对抗训练 之 Projected Gradient Descent（PGD）https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/SmallSampleProblem/AdversarialTraining/AdversarialTraining.md#32-nlp-%E4%B8%AD%E7%BB%8F%E5%85%B8%E5%AF%B9%E6%8A%97%E8%AE%AD%E7%BB%83-%E4%B9%8B--projected-gradient-descentpgd
【关于 脏数据】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning
https://github.com/turned123/NLP-Interview-Notes#52-关于-脏数据那些你不知道的事
【关于 “脏数据”处理】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning
一、动机https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA
1.1 何为“脏数据”？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning/readme.md#11-%E4%BD%95%E4%B8%BA%E8%84%8F%E6%95%B0%E6%8D%AE
1.2 “脏数据” 会带来什么后果？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning/readme.md#12-%E8%84%8F%E6%95%B0%E6%8D%AE-%E4%BC%9A%E5%B8%A6%E6%9D%A5%E4%BB%80%E4%B9%88%E5%90%8E%E6%9E%9C
二、“脏数据” 处理篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning/readme.md#%E4%BA%8C%E8%84%8F%E6%95%B0%E6%8D%AE-%E5%A4%84%E7%90%86%E7%AF%87
2.1 “脏数据” 怎么处理呢？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning/readme.md#21-%E8%84%8F%E6%95%B0%E6%8D%AE-%E6%80%8E%E4%B9%88%E5%A4%84%E7%90%86%E5%91%A2
2.2 置信学习方法篇https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning/readme.md#22-%E7%BD%AE%E4%BF%A1%E5%AD%A6%E4%B9%A0%E6%96%B9%E6%B3%95%E7%AF%87
2.2.1 什么是 置信学习方法？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning/readme.md#221-%E4%BB%80%E4%B9%88%E6%98%AF-%E7%BD%AE%E4%BF%A1%E5%AD%A6%E4%B9%A0%E6%96%B9%E6%B3%95
2.2.2 置信学习方法 优点？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning/readme.md#222-%E7%BD%AE%E4%BF%A1%E5%AD%A6%E4%B9%A0%E6%96%B9%E6%B3%95-%E4%BC%98%E7%82%B9
2.2.3 置信学习方法 怎么做？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning/readme.md#223-%E7%BD%AE%E4%BF%A1%E5%AD%A6%E4%B9%A0%E6%96%B9%E6%B3%95-%E6%80%8E%E4%B9%88%E5%81%9A
2.2.4 置信学习方法 怎么用？有什么开源框架？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning/readme.md#224-%E7%BD%AE%E4%BF%A1%E5%AD%A6%E4%B9%A0%E6%96%B9%E6%B3%95-%E6%80%8E%E4%B9%88%E7%94%A8%E6%9C%89%E4%BB%80%E4%B9%88%E5%BC%80%E6%BA%90%E6%A1%86%E6%9E%B6
2.2.5 置信学习方法 的工作原理？https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/noisy_label_learning/readme.md#225-%E7%BD%AE%E4%BF%A1%E5%AD%A6%E4%B9%A0%E6%96%B9%E6%B3%95-%E7%9A%84%E5%B7%A5%E4%BD%9C%E5%8E%9F%E7%90%86
【关于 炼丹炉】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick
https://github.com/turned123/NLP-Interview-Notes#53-关于-炼丹炉那些你不知道的事
【关于 batch_size设置】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/batch_size
一、训练模型时，batch_size的设置，学习率的设置?https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/batch_size/readme.md#%E4%B8%80%E8%AE%AD%E7%BB%83%E6%A8%A1%E5%9E%8B%E6%97%B6batch_size%E7%9A%84%E8%AE%BE%E7%BD%AE%E5%AD%A6%E4%B9%A0%E7%8E%87%E7%9A%84%E8%AE%BE%E7%BD%AE
【关于 早停法 EarlyStopping 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/EarlyStopping
https://github.com/turned123/NLP-Interview-Notes#54-关于-早停法-earlystopping-那些你不知道的事
【关于 早停法 EarlyStopping 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/EarlyStopping
一、 为什么要用 早停法 EarlyStopping？https://github.com/turned123/NLP-Interview-Notes#%E4%B8%80-%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E7%94%A8-%E6%97%A9%E5%81%9C%E6%B3%95-earlystopping
二、 早停法 EarlyStopping 是什么？https://github.com/turned123/NLP-Interview-Notes#%E4%BA%8C-%E6%97%A9%E5%81%9C%E6%B3%95-earlystopping-%E6%98%AF%E4%BB%80%E4%B9%88
三、早停法 torch 版本怎么实现？https://github.com/turned123/NLP-Interview-Notes#%E4%B8%89%E6%97%A9%E5%81%9C%E6%B3%95-torch-%E7%89%88%E6%9C%AC%E6%80%8E%E4%B9%88%E5%AE%9E%E7%8E%B0
【关于 标签平滑法 LabelSmoothing 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/LabelSmoothing
https://github.com/turned123/NLP-Interview-Notes#55-关于-标签平滑法-labelsmoothing-那些你不知道的事
【关于 标签平滑法 LabelSmoothing 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Trick/LabelSmoothing%EF%BC%9F
一、为什么要有 标签平滑法 LabelSmoothing？https://github.com/turned123/NLP-Interview-Notes#%E4%B8%80%E4%B8%BA%E4%BB%80%E4%B9%88%E8%A6%81%E6%9C%89-%E6%A0%87%E7%AD%BE%E5%B9%B3%E6%BB%91%E6%B3%95-labelsmoothing
二、 标签平滑法 是什么？https://github.com/turned123/NLP-Interview-Notes#%E4%BA%8C-%E6%A0%87%E7%AD%BE%E5%B9%B3%E6%BB%91%E6%B3%95-%E6%98%AF%E4%BB%80%E4%B9%88
三、 标签平滑法 torch 怎么复现？https://github.com/turned123/NLP-Interview-Notes#%E4%B8%89-%E6%A0%87%E7%AD%BE%E5%B9%B3%E6%BB%91%E6%B3%95-torch-%E6%80%8E%E4%B9%88%E5%A4%8D%E7%8E%B0
【关于 Python 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/python
https://github.com/turned123/NLP-Interview-Notes#六关于-python-那些你不知道的事
【关于 Python 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/python
一、什么是*args 和 **kwargs？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#%E4%B8%80%E4%BB%80%E4%B9%88%E6%98%AFargs-%E5%92%8C-kwargs
1.1 为什么会有 *args 和 **kwargs？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#11-%E4%B8%BA%E4%BB%80%E4%B9%88%E4%BC%9A%E6%9C%89-args-%E5%92%8C-kwargs
1.2 *args 和 **kwargs 的用途是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#12-args-%E5%92%8C-kwargs-%E7%9A%84%E7%94%A8%E9%80%94%E6%98%AF%E4%BB%80%E4%B9%88
1.3 *args 是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#13-args-%E6%98%AF%E4%BB%80%E4%B9%88
1.4 **kwargs是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#14-kwargs%E6%98%AF%E4%BB%80%E4%B9%88
1.5 *args 与 **kwargs 的区别是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#15-args-%E4%B8%8E-kwargs-%E7%9A%84%E5%8C%BA%E5%88%AB%E6%98%AF%E4%BB%80%E4%B9%88
二、什么是装饰器？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#%E4%BA%8C%E4%BB%80%E4%B9%88%E6%98%AF%E8%A3%85%E9%A5%B0%E5%99%A8
2.1 装饰器是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#21-%E8%A3%85%E9%A5%B0%E5%99%A8%E6%98%AF%E4%BB%80%E4%B9%88
2.2 装饰器怎么用？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#22-%E8%A3%85%E9%A5%B0%E5%99%A8%E6%80%8E%E4%B9%88%E7%94%A8
三、Python垃圾回收（GC）https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#%E4%B8%89python%E5%9E%83%E5%9C%BE%E5%9B%9E%E6%94%B6gc
3.1 垃圾回收算法有哪些？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#31-%E5%9E%83%E5%9C%BE%E5%9B%9E%E6%94%B6%E7%AE%97%E6%B3%95%E6%9C%89%E5%93%AA%E4%BA%9B
3.2 引用计数（主要）是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#32-%E5%BC%95%E7%94%A8%E8%AE%A1%E6%95%B0%E4%B8%BB%E8%A6%81%E6%98%AF%E4%BB%80%E4%B9%88
3.3 标记-清除是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#33-%E6%A0%87%E8%AE%B0-%E6%B8%85%E9%99%A4%E6%98%AF%E4%BB%80%E4%B9%88
3.4 分代回收是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#34-%E5%88%86%E4%BB%A3%E5%9B%9E%E6%94%B6%E6%98%AF%E4%BB%80%E4%B9%88
四、python的sorted函数对字典按key排序和按value排序https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#%E5%9B%9Bpython%E7%9A%84sorted%E5%87%BD%E6%95%B0%E5%AF%B9%E5%AD%97%E5%85%B8%E6%8C%89key%E6%8E%92%E5%BA%8F%E5%92%8C%E6%8C%89value%E6%8E%92%E5%BA%8F
4.1 python 的sorted函数是什么？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#41-python-%E7%9A%84sorted%E5%87%BD%E6%95%B0%E6%98%AF%E4%BB%80%E4%B9%88
4.2 python 的sorted函数举例说明？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#42-python-%E7%9A%84sorted%E5%87%BD%E6%95%B0%E4%B8%BE%E4%BE%8B%E8%AF%B4%E6%98%8E
五、直接赋值、浅拷贝和深度拷贝https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#%E4%BA%94%E7%9B%B4%E6%8E%A5%E8%B5%8B%E5%80%BC%E6%B5%85%E6%8B%B7%E8%B4%9D%E5%92%8C%E6%B7%B1%E5%BA%A6%E6%8B%B7%E8%B4%9D
5.1 概念介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#51-%E6%A6%82%E5%BF%B5%E4%BB%8B%E7%BB%8D
5.2 介绍https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#52-%E4%BB%8B%E7%BB%8D
5.3 变量定义流程https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#53-%E5%8F%98%E9%87%8F%E5%AE%9A%E4%B9%89%E6%B5%81%E7%A8%8B
5.3 赋值https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#53-%E8%B5%8B%E5%80%BC
5.4 浅拷贝https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#54-%E6%B5%85%E6%8B%B7%E8%B4%9D
5.5 深度拷贝https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#55--%E6%B7%B1%E5%BA%A6%E6%8B%B7%E8%B4%9D
5.6 核心：不可变对象类型 and 可变对象类型https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#56-%E6%A0%B8%E5%BF%83%E4%B8%8D%E5%8F%AF%E5%8F%98%E5%AF%B9%E8%B1%A1%E7%B1%BB%E5%9E%8B-and-%E5%8F%AF%E5%8F%98%E5%AF%B9%E8%B1%A1%E7%B1%BB%E5%9E%8B
5.6.1 不可变对象类型https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#561-%E4%B8%8D%E5%8F%AF%E5%8F%98%E5%AF%B9%E8%B1%A1%E7%B1%BB%E5%9E%8B
5.6.2 可变对象类型https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#562-%E5%8F%AF%E5%8F%98%E5%AF%B9%E8%B1%A1%E7%B1%BB%E5%9E%8B
六、进程、线程、协程https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#%E5%85%AD%E8%BF%9B%E7%A8%8B%E7%BA%BF%E7%A8%8B%E5%8D%8F%E7%A8%8B
6.1 进程https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#61-%E8%BF%9B%E7%A8%8B
6.1.1 什么是进程？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#611-%E4%BB%80%E4%B9%88%E6%98%AF%E8%BF%9B%E7%A8%8B
6.1.2 进程间如何通信？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#612-%E8%BF%9B%E7%A8%8B%E9%97%B4%E5%A6%82%E4%BD%95%E9%80%9A%E4%BF%A1
6.2 线程https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#62-%E7%BA%BF%E7%A8%8B
6.2.1 什么是线程？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#621-%E4%BB%80%E4%B9%88%E6%98%AF%E7%BA%BF%E7%A8%8B
6.2.2 线程间如何通信？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#622-%E7%BA%BF%E7%A8%8B%E9%97%B4%E5%A6%82%E4%BD%95%E9%80%9A%E4%BF%A1
6.3 进程 vs 线程https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#63-%E8%BF%9B%E7%A8%8B-vs-%E7%BA%BF%E7%A8%8B
6.3.1 区别https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#631-%E5%8C%BA%E5%88%AB
6.3.2 应用场景https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#632-%E5%BA%94%E7%94%A8%E5%9C%BA%E6%99%AF
6.4 协程https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#64-%E5%8D%8F%E7%A8%8B
6.4.1 什么是协程？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#641-%E4%BB%80%E4%B9%88%E6%98%AF%E5%8D%8F%E7%A8%8B
6.4.2 协程的优点？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#642-%E5%8D%8F%E7%A8%8B%E7%9A%84%E4%BC%98%E7%82%B9
七、全局解释器锁https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#%E4%B8%83%E5%85%A8%E5%B1%80%E8%A7%A3%E9%87%8A%E5%99%A8%E9%94%81
7.1 什么是全局解释器锁？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#71-%E4%BB%80%E4%B9%88%E6%98%AF%E5%85%A8%E5%B1%80%E8%A7%A3%E9%87%8A%E5%99%A8%E9%94%81
7.2 GIL有什么作用？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#72-gil%E6%9C%89%E4%BB%80%E4%B9%88%E4%BD%9C%E7%94%A8
7.3 GIL有什么影响？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#73-gil%E6%9C%89%E4%BB%80%E4%B9%88%E5%BD%B1%E5%93%8D
7.4 如何避免GIL带来的影响？https://github.com/turned123/NLP-Interview-Notes/blob/main/python/readme.md#74-%E5%A6%82%E4%BD%95%E9%81%BF%E5%85%8Dgil%E5%B8%A6%E6%9D%A5%E7%9A%84%E5%BD%B1%E5%93%8D
【关于 Tensorflow 】那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow
https://github.com/turned123/NLP-Interview-Notes#七关于-tensorflow-那些你不知道的事
【关于 Tensorflow 损失函数】 那些你不知道的事https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study
一、动机https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#%E4%B8%80%E5%8A%A8%E6%9C%BA
二、什么是损失函数？https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#%E4%BA%8C%E4%BB%80%E4%B9%88%E6%98%AF%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0
三、目标函数、损失函数、代价函数之间的关系与区别？https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#%E4%B8%89%E7%9B%AE%E6%A0%87%E5%87%BD%E6%95%B0%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0%E4%BB%A3%E4%BB%B7%E5%87%BD%E6%95%B0%E4%B9%8B%E9%97%B4%E7%9A%84%E5%85%B3%E7%B3%BB%E4%B8%8E%E5%8C%BA%E5%88%AB
四、损失函数的类别https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#%E5%9B%9B%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0%E7%9A%84%E7%B1%BB%E5%88%AB
4.1 回归模型的损失函数https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#41-%E5%9B%9E%E5%BD%92%E6%A8%A1%E5%9E%8B%E7%9A%84%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0
（1）L1正则损失函数（即绝对值损失函数）https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#1l1%E6%AD%A3%E5%88%99%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0%E5%8D%B3%E7%BB%9D%E5%AF%B9%E5%80%BC%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0
（2）L2正则损失函数（即欧拉损失函数）https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#2l2%E6%AD%A3%E5%88%99%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0%E5%8D%B3%E6%AC%A7%E6%8B%89%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0
（3）均方误差（MSE, mean squared error）https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#3%E5%9D%87%E6%96%B9%E8%AF%AF%E5%B7%AEmse-mean-squared-error
（4）Pseudo-Huber 损失函数https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#4pseudo-huber-%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0
4.2 分类模型的损失函数https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#42-%E5%88%86%E7%B1%BB%E6%A8%A1%E5%9E%8B%E7%9A%84%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0
（1）Hinge损失函数https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#1hinge%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0
（2）两类交叉熵（Cross-entropy）损失函数https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#2%E4%B8%A4%E7%B1%BB%E4%BA%A4%E5%8F%89%E7%86%B5cross-entropy%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0
（3）Sigmoid交叉熵损失函数https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#3sigmoid%E4%BA%A4%E5%8F%89%E7%86%B5%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0
（4）加权交叉熵损失函数https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#4%E5%8A%A0%E6%9D%83%E4%BA%A4%E5%8F%89%E7%86%B5%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0
（5）Softmax交叉熵损失函数https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#5softmax%E4%BA%A4%E5%8F%89%E7%86%B5%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0
(6) SparseCategoricalCrossentropy vs sparse_categorical_crossentropyhttps://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#6-sparsecategoricalcrossentropy-vs-sparse_categorical_crossentropy
五、总结https://github.com/turned123/NLP-Interview-Notes/blob/main/Tensorflow/loss_study/readme.md#%E4%BA%94%E6%80%BB%E7%BB%93
Readme https://github.com/turned123/NLP-Interview-Notes#readme-ov-file
Please reload this pagehttps://github.com/turned123/NLP-Interview-Notes
Activityhttps://github.com/turned123/NLP-Interview-Notes/activity
0 forkshttps://github.com/turned123/NLP-Interview-Notes/forks
Report repository https://github.com/contact/report-content?content_url=https%3A%2F%2Fgithub.com%2Fturned123%2FNLP-Interview-Notes&report=turned123+%28user%29
Releaseshttps://github.com/turned123/NLP-Interview-Notes/releases
Packages 0https://github.com/users/turned123/packages?repo_name=NLP-Interview-Notes
Please reload this pagehttps://github.com/turned123/NLP-Interview-Notes
Contributorshttps://github.com/turned123/NLP-Interview-Notes/graphs/contributors
Please reload this pagehttps://github.com/turned123/NLP-Interview-Notes
https://github.com
Termshttps://docs.github.com/site-policy/github-terms/github-terms-of-service
Privacyhttps://docs.github.com/site-policy/privacy-policies/github-privacy-statement
Securityhttps://github.com/security
Statushttps://www.githubstatus.com/
Communityhttps://github.community/
Docshttps://docs.github.com/
Contacthttps://support.github.com?tags=dotcom-footer

Viewport: width=device-width

URLs of crawlers that visited me.