{"id":"W4319862706","doi":"10.1109/slt54892.2023.10022986","title":"Flow-ER: A Flow-Based Embedding Regularization Strategy for Robust Speech Representation Learning","year":2023,"lang":"en","type":"article","venue":"2022 IEEE Spoken Language Technology Workshop (SLT)","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Computer Research Institute of Montréal","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Embedding; Computer science; Regularization (linguistics); Artificial intelligence; Machine learning; Deep learning; Speech recognition; Representation (politics); Feature learning; Bottleneck; Task (project management)","routes":{"ca_aff":true,"ca_fund":true,"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":[0.001766751,0.001562814,0.0009916844,0.0009059579,0.000369178,0.000587513,0.001751122,0.001515736,0.003690646],"category_scores_gemma":[0.002407772,0.0005775217,0.0009811326,0.0006532223,0.0007109183,0.001812455,0.001554761,0.001999963,0.00143916],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005441116,"about_ca_system_score_gemma":0.0008777839,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002373785,"about_ca_topic_score_gemma":0.002319861,"domain_scores_codex":[0.9994683,0.0001602654,0.00003020086,0.0001322475,0.0001573799,0.00005172311],"domain_scores_gemma":[0.999494,0.0002113935,0.00005304901,0.00007250546,0.0001358342,0.00003328525],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0002575577,0.0002516914,0.0004148791,0.0002220195,0.0001225506,0.000138845,0.0001396745,0.2915009,0.04360294,0.03294125,0.01148109,0.6189266],"study_design_scores_gemma":[0.000009179236,0.00004177257,0.00005640015,0.000006912397,0.000006834458,0.00003369824,0.000004007248,0.9908171,0.003893957,0.003460849,0.001658717,0.00001061019],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.001545307,0.0001277622,0.9972474,0.00005828358,0.00002771448,0.00003136296,0.00003263895,0.0004979027,0.0004315835],"genre_scores_gemma":[0.1428676,0.000643771,0.8449152,0.0004043699,0.0001821278,0.0003575829,0.0007751561,0.0006240839,0.009230142],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003690646,"threshold_uncertainty_score":0.01234651,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03353607308453113,"score_gpt":0.30544206453765,"score_spread":0.2719059914531189,"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."}}