{"id":"W4408354432","doi":"10.1109/icassp49660.2025.10889092","title":"What Are They Doing? Joint Audio-Speech Co-Reasoning","year":2025,"lang":"en","type":"article","venue":"","topic":"Speech and dialogue systems","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute; Concordia University","funders":"","keywords":"Joint (building); Computer science; Speech recognition; Engineering","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":[],"consensus_categories":[],"category_scores_codex":[0.0004437216,0.0001438648,0.0002200997,0.0001466674,0.0001483237,0.0008188165,0.0006593392,0.0000754735,0.00003477579],"category_scores_gemma":[0.00007199867,0.0001147618,0.00009442391,0.0003073969,0.0000209466,0.000779382,0.0002273293,0.0001141997,0.0004437547],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006333506,"about_ca_system_score_gemma":0.0001026784,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002618751,"about_ca_topic_score_gemma":0.0002134286,"domain_scores_codex":[0.9987017,0.00007344373,0.00023482,0.0004056476,0.0002495432,0.000334855],"domain_scores_gemma":[0.998956,0.00008214733,0.0000828122,0.0006929183,0.00008978407,0.00009635866],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00002557504,0.0002157683,0.02038611,0.0001716862,0.00017219,0.0002815791,0.002863185,0.0001307674,0.005648101,0.5670058,0.1815711,0.2215281],"study_design_scores_gemma":[0.005477847,0.0003441179,0.06478722,0.004706474,0.0000615432,0.0004078945,0.008001537,0.05211102,0.3176219,0.06121297,0.4822047,0.003062754],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02483152,0.001204976,0.5922024,0.005292979,0.006307469,0.0004552091,0.000002291729,0.001061291,0.3686419],"genre_scores_gemma":[0.9412035,0.00005701051,0.04297673,0.00269918,0.0002015805,0.00002768433,0.000003905307,0.00001079669,0.01281967],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9163719,"threshold_uncertainty_score":0.7895866,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01509403267989033,"score_gpt":0.2614817200377479,"score_spread":0.2463876873578576,"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."}}