{"id":"W4317383069","doi":"10.1109/robio55434.2022.10011692","title":"Real-time Architecture for Audio-Visual Active Speaker Detection","year":2022,"lang":"en","type":"article","venue":"2022 IEEE International Conference on Robotics and Biomimetics (ROBIO)","topic":"Speech and Audio Processing","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"","keywords":"Computer science; Robot; Frame (networking); Code (set theory); Architecture; Artificial intelligence; Frame rate; Speech recognition; Quality (philosophy)","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":[0.0007571176,0.001040626,0.0007045268,0.0007113043,0.0003533303,0.0007628854,0.00261313,0.0008850224,0.00570047],"category_scores_gemma":[0.00199401,0.0004409639,0.0006046921,0.0004182417,0.0003705012,0.001206614,0.001170523,0.001325299,0.003617486],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008898333,"about_ca_system_score_gemma":0.001052548,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005852851,"about_ca_topic_score_gemma":0.008034175,"domain_scores_codex":[0.9994106,0.00004875506,0.00003100911,0.0002540796,0.0001748857,0.0000806371],"domain_scores_gemma":[0.9993906,0.0001365854,0.00004506616,0.0001068062,0.0002515236,0.00006943588],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001391276,0.0004770374,0.002565324,0.0002804369,0.0001798691,0.0001843403,0.0001862005,0.0515324,0.09933952,0.003257063,0.01939881,0.8212076],"study_design_scores_gemma":[0.00006528301,0.0002793417,0.001413964,0.00002168947,0.00007864173,0.0002318699,0.00003200355,0.9467677,0.03857018,0.00229775,0.01019193,0.00004964916],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01895297,0.0006825682,0.9516961,0.0001747309,0.000299035,0.0001610085,0.0004033623,0.02510753,0.002522718],"genre_scores_gemma":[0.5530743,0.0003607496,0.4301689,0.0004762814,0.00022167,0.000545057,0.002069961,0.0008325041,0.01225065],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005852851,"threshold_uncertainty_score":0.01906997,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02706687756412575,"score_gpt":0.2892053514631431,"score_spread":0.2621384738990174,"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."}}