{"id":"W4386273179","doi":"10.2139/ssrn.4733627","title":"Speech Self-Supervised Representations Benchmarking: A Case for Larger Probing Heads","year":2024,"lang":"en","type":"preprint","venue":"SSRN Electronic Journal","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"Concordia University","funders":"","keywords":"Benchmarking; Computer science; Ranking (information retrieval); Inference; Downstream (manufacturing); Task (project management); Generalization; Feature (linguistics); Set (abstract data type); Artificial intelligence; Architecture; Machine learning; Natural language processing; Engineering; Mathematics; Geography","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":["metaepi_narrow","scholarly_communication","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.003245212,0.0004045599,0.0004298006,0.0005506675,0.0005064543,0.001187627,0.001076354,0.0003066954,0.00006043168],"category_scores_gemma":[0.0001396631,0.0003752591,0.0006420938,0.0004295187,0.00002684144,0.0002620103,0.0008045622,0.004103676,0.00007460984],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001309036,"about_ca_system_score_gemma":0.005549154,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007380334,"about_ca_topic_score_gemma":0.000454091,"domain_scores_codex":[0.995367,0.0002148746,0.0006811845,0.0009159668,0.0004653223,0.002355657],"domain_scores_gemma":[0.9981878,0.0002641956,0.0002935343,0.0007083644,0.0003524421,0.000193614],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00003424321,0.0004010771,0.00006008533,0.0003948789,0.001789669,0.002143058,0.002856152,0.0001040268,0.0003465644,0.224581,0.002958432,0.7643307],"study_design_scores_gemma":[0.0006439643,0.0001578625,0.000005781428,0.000254255,0.0002664925,0.0383865,0.0008136501,0.1290011,0.0006569182,0.8265406,0.002592251,0.0006805774],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06089819,0.006364606,0.9106826,0.01028659,0.004570647,0.002159368,0.00005931241,0.001000197,0.003978561],"genre_scores_gemma":[0.5216573,0.004797571,0.464718,0.0006033866,0.003997791,0.0007084171,0.00006503622,0.0002109943,0.003241474],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7636502,"threshold_uncertainty_score":0.9998699,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02543416354791751,"score_gpt":0.2923444770256426,"score_spread":0.2669103134777251,"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."}}