{"id":"W4319299923","doi":"10.1109/wacv56688.2023.00224","title":"AudioViewer: Learning to Visualize Sounds","year":2023,"lang":"en","type":"article","venue":"2023 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)","topic":"Music and Audio Processing","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"","keywords":"Computer science; Speech recognition; Perception; Field (mathematics); Visualization; Substitution (logic); Natural language processing; Parsing; Artificial intelligence; Human–computer interaction; Multimedia","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":[0.0005829387,0.001589967,0.0004803574,0.0006939483,0.0002540064,0.001013294,0.002049821,0.001072214,0.01723476],"category_scores_gemma":[0.003509238,0.0004580257,0.0008648801,0.0003575151,0.0005063354,0.0023871,0.002133305,0.001408328,0.006154065],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003950576,"about_ca_system_score_gemma":0.0004066397,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001682637,"about_ca_topic_score_gemma":0.003401746,"domain_scores_codex":[0.9996161,0.00008327845,0.0000112509,0.000178724,0.00008067438,0.00002995712],"domain_scores_gemma":[0.9993728,0.0003122987,0.0000315887,0.0001250302,0.00009119389,0.0000672069],"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":[0.0004769884,0.0002210334,0.001318863,0.0004267424,0.00008446801,0.0002018533,0.0002757539,0.02869305,0.03552578,0.007859519,0.06264543,0.8622705],"study_design_scores_gemma":[0.0001517393,0.0004393485,0.001510192,0.0001542438,0.00005120187,0.0004611512,0.0002671114,0.8448673,0.05033246,0.03451629,0.0671654,0.00008356765],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02476118,0.0009309408,0.8988876,0.0004225464,0.0003475464,0.0002443574,0.00386787,0.06133208,0.009205897],"genre_scores_gemma":[0.1847645,0.0009044745,0.7834331,0.0004853253,0.0001434258,0.0006999574,0.01027229,0.003120969,0.01617581],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01723476,"threshold_uncertainty_score":0.05765593,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03231023235635655,"score_gpt":0.3323631389603665,"score_spread":0.30005290660401,"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."}}