{"id":"W4372260310","doi":"10.1109/icassp49357.2023.10095969","title":"Large-Scale Contrastive Language-Audio Pretraining with Feature Fusion and Keyword-to-Caption Augmentation","year":2023,"lang":"en","type":"article","venue":"","topic":"Music and Audio Processing","field":"Computer Science","cited_by":385,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Computer science; Audio mining; Natural language processing; Pipeline (software); Speech recognition; Artificial intelligence; Encoder; Construct (python library); Feature (linguistics); Natural language; Language model; Speech processing; Acoustic model; Linguistics","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":[0.001270851,0.001743469,0.0009132901,0.0008115565,0.0004101219,0.0006636875,0.002643388,0.001437339,0.005466122],"category_scores_gemma":[0.003837546,0.0005868577,0.001310511,0.000704349,0.0007602539,0.002048927,0.001948971,0.002646201,0.002889117],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009679787,"about_ca_system_score_gemma":0.001084727,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005359553,"about_ca_topic_score_gemma":0.008192667,"domain_scores_codex":[0.9992831,0.0001629465,0.00003141254,0.0002862998,0.0001428941,0.00009319477],"domain_scores_gemma":[0.9986265,0.0007540575,0.00007450895,0.000191252,0.0002820877,0.00007158618],"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":[0.0007685081,0.0006944359,0.001629032,0.0002777838,0.0001740924,0.0002570602,0.0001840742,0.1224904,0.06912414,0.002621716,0.01254507,0.7892338],"study_design_scores_gemma":[0.00003490027,0.0002243682,0.0005227529,0.00001760269,0.00003865797,0.0000894573,0.00004240269,0.9641606,0.03067479,0.001871727,0.002298339,0.00002446863],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05716269,0.0008006398,0.92123,0.0003682797,0.0002041539,0.0002828289,0.0006759436,0.01492501,0.004350474],"genre_scores_gemma":[0.5205704,0.0003439093,0.4600874,0.000974104,0.0001392692,0.0007220655,0.004517694,0.0007113089,0.01193375],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005466122,"threshold_uncertainty_score":0.01828605,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.00899152659473625,"score_gpt":0.2488691384566329,"score_spread":0.2398776118618966,"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."}}