{"id":"W4388235576","doi":"10.1109/sampta59647.2023.10301400","title":"Sampling Informative Positives Pairs in Contrastive Learning","year":2023,"lang":"en","type":"article","venue":"","topic":"Machine Learning and Algorithms","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Similarity (geometry); Sampling (signal processing); Artificial intelligence; Representation (politics); Computer science; Class (philosophy); Space (punctuation); False positive paradox; Machine learning; Mathematics; Pattern recognition (psychology)","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":[0.000438744,0.00008892568,0.0001155874,0.0002160105,0.0001163035,0.000107551,0.0002559991,0.00002960461,0.00001602494],"category_scores_gemma":[0.0001718723,0.00007607169,0.00003132825,0.0007681311,0.00002166316,0.0004468968,0.0001746263,0.0002979784,0.000259475],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002072153,"about_ca_system_score_gemma":0.00002608505,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009422051,"about_ca_topic_score_gemma":0.00000720083,"domain_scores_codex":[0.9991668,0.00008648141,0.0001559147,0.0001762652,0.000145194,0.0002692931],"domain_scores_gemma":[0.999419,0.0003582195,0.00004645249,0.00009578371,0.00003295657,0.00004761719],"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":[0.00001782988,0.00005977906,0.07248076,0.00003607464,0.0000526076,0.0001035546,0.05782646,0.1256834,0.000404458,0.1365779,0.0004944609,0.6062627],"study_design_scores_gemma":[0.000305494,0.0000727381,0.1295332,0.00003615125,7.155515e-7,0.000004517722,0.001816342,0.8658621,0.0001801946,0.001074679,0.0009635114,0.0001502916],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.09390523,0.00001747009,0.8736685,0.001107134,0.0001680054,0.000111664,6.752905e-7,0.0009935854,0.03002773],"genre_scores_gemma":[0.9834825,0.000008188264,0.01495903,0.0001558155,0.00002927558,0.000008882413,0.000004875999,0.000004967505,0.001346477],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8895773,"threshold_uncertainty_score":0.3335113,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01514522349937,"score_gpt":0.2783324649142953,"score_spread":0.2631872414149253,"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."}}