{"id":"W3143367551","doi":"10.21437/interspeech.2021-941","title":"ECAPA-TDNN Embeddings for Speaker Diarization","year":2021,"lang":"en","type":"article","venue":"","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":92,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Sherbrooke; Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Speaker diarisation; Computer science; Speaker recognition; Robustness (evolution); Speech recognition; Artificial neural network; Discriminative model; Artificial intelligence; Time delay neural network; Pattern recognition (psychology)","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":[0.000815652,0.0009922698,0.0005938901,0.0004560119,0.0003617363,0.0006819298,0.001245114,0.0009348334,0.005347266],"category_scores_gemma":[0.002531846,0.0003464745,0.0006225125,0.0005967905,0.0003472061,0.001396662,0.001492647,0.002285283,0.00524782],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000459664,"about_ca_system_score_gemma":0.0006355228,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003129293,"about_ca_topic_score_gemma":0.005445018,"domain_scores_codex":[0.9992552,0.0001329216,0.00004559478,0.0002804015,0.0002249503,0.0000608985],"domain_scores_gemma":[0.9992736,0.0001634964,0.00004858472,0.0002296805,0.0002437758,0.00004092145],"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":[0.0005212355,0.0001963946,0.001006828,0.0001639582,0.0001320802,0.0001517473,0.0001343164,0.08342674,0.06705797,0.005923178,0.01614712,0.8251384],"study_design_scores_gemma":[0.00001401964,0.00006813258,0.0007275295,0.0000133542,0.00002361868,0.0002101916,0.00002428251,0.962737,0.02500732,0.003813287,0.007333625,0.00002763058],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02041622,0.0008121345,0.9676846,0.0002060351,0.0004098669,0.00006362372,0.0006808893,0.006034779,0.003691834],"genre_scores_gemma":[0.4274446,0.0007655724,0.5448315,0.0004472282,0.0002265969,0.0002742749,0.00525278,0.001057886,0.01969964],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005347266,"threshold_uncertainty_score":0.01788837,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01856822171074096,"score_gpt":0.2489785712134402,"score_spread":0.2304103495026992,"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."}}