René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Weighting factor for positive examples (conf['data']['positive_example_penalty']) only works for 'maxhinge' target","articleBody":"The only time in which `positive_example_penalty` appears in the codebase is in:\r\nhttps://github.com/PPPLDeepLearning/plasma-python/blob/c82ba61e339882a5af10b1052edc0348e16119f4/plasma/conf_parser.py#L86-L102\r\nwhich is loaded only for an unused method in the `MaxHingeTarget` class, noted here:\r\nhttps://github.com/PPPLDeepLearning/plasma-python/blob/7986f468e43a56a5ae845dd1b88cf9fca048ac5a/plasma/models/targets.py#L153\r\n\r\nNeed to extend it to the other target functions, remove it as a parameter, or document this more thoroughly. \r\n\r\nNot as important for DIII-D datasets as it is for our JET datasets (for which the non-/disruptive classes are more imbalanced).\r\n\r\n- [ ] Also, `conf['training']['ranking_difficulty_fac']: 1.0 # how much to upweight incorrectly classified shots during training` appears to perform a very related role, but instead within `loader.py`, `mpi_runner.py`, and `performance.py`.","author":{"url":"https://github.com/felker","@type":"Person","name":"felker"},"datePublished":"2019-12-19T16:51:37.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":0},"url":"https://github.com/53/plasma-python/issues/53"}