{"id":"W2935542736","doi":"10.21437/interspeech.2019-2605","title":"Learning Problem-Agnostic Speech Representations from Multiple Self-Supervised Tasks","year":2019,"lang":"en","type":"preprint","venue":"","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":34,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Compute Canada","keywords":"Computer science; Encoder; Discriminator; Artificial intelligence; Speech recognition; Machine learning; Supervised learning; Construct (python library); SIGNAL (programming language); Adaptation (eye); Identity (music); Pattern recognition (psychology); Artificial neural network; Psychology","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.0003324169,0.0003916063,0.0004821144,0.0002688771,0.000183341,0.0007017441,0.001600456,0.0003752374,0.001295495],"category_scores_gemma":[0.0006210298,0.0003781666,0.0002838081,0.0002865305,0.00002853371,0.0003340806,0.00183908,0.0009347132,0.003061553],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009593952,"about_ca_system_score_gemma":0.0002615809,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001587923,"about_ca_topic_score_gemma":0.000115998,"domain_scores_codex":[0.9967805,0.0003313164,0.0005541253,0.001340474,0.000586428,0.0004071704],"domain_scores_gemma":[0.9961324,0.001671908,0.0002599275,0.001484514,0.000268642,0.0001826552],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00003651322,0.001472008,0.07378094,0.0005239363,0.001272509,0.0003415998,0.009942056,0.01494979,0.003466003,0.00260615,0.02369642,0.8679121],"study_design_scores_gemma":[0.0009860878,0.00005910059,0.006827567,0.000265339,0.00009584768,0.00001863969,0.0002900903,0.9665409,0.008543901,0.01053775,0.004807892,0.001026866],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.08527295,0.0001513079,0.7812846,0.002547338,0.002205865,0.002282565,0.00008197597,0.00344127,0.1227322],"genre_scores_gemma":[0.2394414,0.0001149851,0.754972,0.0003886157,0.0001847025,0.0001590814,0.0002299154,0.00004089295,0.004468369],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9515911,"threshold_uncertainty_score":0.999867,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03083472838545893,"score_gpt":0.2633205265083643,"score_spread":0.2324857981229054,"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."}}