{"id":"W4417068938","doi":"10.1109/iccv51701.2025.02126","title":"Aurelia: Test-Time Reasoning Distillation in Audio-Visual LLMs","year":2025,"lang":"en","type":"preprint","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology; University of Toronto","funders":"","keywords":"Benchmark (surveying); Process (computing); Case-based reasoning; Visual reasoning; Reasoning system; Deductive reasoning; Code (set theory)","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":[0.002897703,0.00232258,0.0009365143,0.000813422,0.000574537,0.002346424,0.00422746,0.0028547,0.01824613],"category_scores_gemma":[0.01798454,0.0006528738,0.00142435,0.0004662271,0.001267279,0.003235309,0.003126492,0.0035693,0.004125195],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001919648,"about_ca_system_score_gemma":0.002239556,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01355899,"about_ca_topic_score_gemma":0.01984696,"domain_scores_codex":[0.9975055,0.0008892804,0.0001404471,0.0007648777,0.0004650757,0.0002348435],"domain_scores_gemma":[0.9945415,0.003931959,0.0001612377,0.0006307242,0.0005041535,0.0002305681],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001711207,0.0006932453,0.003892013,0.001635542,0.0003139042,0.000718792,0.0005042524,0.4550811,0.01854201,0.01934667,0.07020345,0.4273578],"study_design_scores_gemma":[0.0001956454,0.0001058737,0.0002719917,0.000048551,0.00001872883,0.00006083321,0.00007758561,0.9748489,0.006860089,0.01023463,0.007255022,0.00002209035],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.09165498,0.002661303,0.7215727,0.002109925,0.001008848,0.0008733981,0.00740177,0.1477991,0.02491796],"genre_scores_gemma":[0.5157709,0.000314588,0.4577376,0.001473165,0.0001295499,0.0006677603,0.011451,0.003736349,0.008719089],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01824613,"threshold_uncertainty_score":0.06103945,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.006768323365754613,"score_gpt":0.2952767481250392,"score_spread":0.2885084247592846,"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."}}