{"id":"W4391901264","doi":"10.48550/arxiv.2402.10150","title":"$f$-MICL: Understanding and Generalizing InfoNCE-based Contrastive Learning","year":2024,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Fuzzy Logic and Control Systems","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Government of Canada; Canadian Institute for Advanced Research","keywords":"Contrastive analysis; Psychology; Linguistics; Computer science; Natural language processing; Artificial intelligence; Mathematics education; Philosophy","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005133824,0.001479555,0.001339867,0.001889518,0.0007989327,0.002046834,0.002701214,0.002760664,0.001688294],"category_scores_gemma":[0.01541351,0.0004586664,0.001012138,0.001159583,0.002896499,0.004934744,0.00437584,0.003152705,0.0005973906],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002159999,"about_ca_system_score_gemma":0.001178618,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002364814,"about_ca_topic_score_gemma":0.002915435,"domain_scores_codex":[0.9973901,0.001050712,0.0001298994,0.0006611898,0.0006085714,0.0001595694],"domain_scores_gemma":[0.9948606,0.002982318,0.0003783199,0.0008310043,0.0007561137,0.0001915509],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003832023,0.0003380335,0.005323329,0.0004481433,0.000243257,0.000218665,0.0004756441,0.3844703,0.01227391,0.09978507,0.009038378,0.4870021],"study_design_scores_gemma":[0.00002179388,0.000150448,0.0005127944,0.00002139041,0.00001748874,0.0000794463,0.00003258141,0.9303614,0.002885495,0.06427247,0.001621334,0.00002333103],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02359582,0.000642705,0.9714433,0.0005788404,0.00004662902,0.00008053124,0.0001259281,0.0009831534,0.002503229],"genre_scores_gemma":[0.6456343,0.0004393688,0.3476642,0.001530128,0.0002691708,0.0003281549,0.0008307324,0.0004825446,0.002821468],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005133824,"threshold_uncertainty_score":0.02715057,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1102722430120212,"score_gpt":0.1780003741759273,"score_spread":0.06772813116390604,"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."}}