{"id":"W3106085430","doi":"","title":"Gradient Estimation with Stochastic Softmax Tricks","year":2020,"lang":"en","type":"article","venue":"Neural Information Processing Systems","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Softmax function; Estimator; Gumbel distribution; Computer science; Latent variable; Algorithm; Artificial intelligence; Mathematics; Theoretical computer science; Machine learning; Artificial neural network; Statistics","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003373124,0.0015653,0.001727736,0.0009471912,0.000474812,0.001677649,0.002151024,0.001542135,0.005383845],"category_scores_gemma":[0.01678122,0.001088697,0.001284556,0.001250748,0.001435512,0.003190086,0.002337893,0.004113864,0.002165195],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009272807,"about_ca_system_score_gemma":0.001472938,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001710096,"about_ca_topic_score_gemma":0.001977395,"domain_scores_codex":[0.9979965,0.001042839,0.0001082577,0.0003363359,0.0003945195,0.0001215711],"domain_scores_gemma":[0.9970293,0.001870097,0.0002304342,0.000456436,0.0003157605,0.00009785153],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001903225,0.0001001853,0.000616848,0.0002267643,0.0001578644,0.00009184835,0.000104906,0.656009,0.003496607,0.1910324,0.008451246,0.1395219],"study_design_scores_gemma":[0.00001251006,0.00001457848,0.00005276767,0.00001048871,0.000006537502,0.00001182528,0.000003780719,0.9605463,0.0004593112,0.03811477,0.0007593507,0.000007827894],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001671651,0.0001112286,0.997043,0.0001736171,0.00002754682,0.0000214345,0.0000428378,0.0003142543,0.0005944925],"genre_scores_gemma":[0.2459655,0.0007942143,0.743158,0.0007646433,0.0003356629,0.0005696972,0.0007738659,0.0006374295,0.007001031],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005383845,"threshold_uncertainty_score":0.0180108,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01854419720311598,"score_gpt":0.2238023871480692,"score_spread":0.2052581899449533,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}