{"id":"W3174726724","doi":"10.18653/v1/2021.acl-long.163","title":"Optimizing Deeper Transformers on Small Datasets","year":2021,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":57,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute; McGill University; Canadian Institute for Advanced Research; University of Alberta; University of Waterloo","funders":"","keywords":"Computer science; Computational linguistics; Natural language processing; Transformer; Artificial intelligence; Library science; Cognitive science; Engineering; Psychology; Electrical engineering","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00008784866,0.00006662749,0.00006815996,0.0000279348,0.0000516532,0.0000879206,0.0003224961,0.00002709267,0.00008854784],"category_scores_gemma":[0.00001074421,0.00005901996,0.00003365079,0.0001116139,0.000006065201,0.0001839062,0.00007626795,0.00007686952,0.00006602568],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001645248,"about_ca_system_score_gemma":0.00004672265,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001611268,"about_ca_topic_score_gemma":0.00003984408,"domain_scores_codex":[0.9992942,0.00001882746,0.0001029759,0.000290401,0.000119153,0.0001744993],"domain_scores_gemma":[0.9994765,0.00003322326,0.00001016778,0.0004012948,0.00001824787,0.00006056522],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000006033988,0.0001527093,0.00007587891,0.00003065233,0.00004315622,0.0003199695,0.001993785,0.02357994,0.003256093,0.5046585,0.003069479,0.4628138],"study_design_scores_gemma":[0.00087007,0.00007129595,0.0001071823,0.00005269878,0.00001054995,0.00007418593,0.0002385636,0.8570588,0.06897847,0.003555171,0.0684271,0.0005559214],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003096481,0.00004193115,0.975207,0.001598447,0.0001847838,0.00003330415,0.000003273009,0.00009066382,0.01974415],"genre_scores_gemma":[0.1525608,0.00001954532,0.8426746,0.003865635,0.00004770082,0.000004564143,0.00002445536,0.000007166126,0.0007955606],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.8334789,"threshold_uncertainty_score":0.2406764,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03922932592945792,"score_gpt":0.2517262054989651,"score_spread":0.2124968795695072,"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."}}