{"id":"W4416136027","doi":"10.48550/arxiv.2506.05027","title":"Tuning the Right Foundation Models is What you Need for Partial Label Learning","year":2025,"lang":"en","type":"preprint","venue":"ArXiv.org","topic":"Text and Document Classification Technologies","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Canadian Institute for Advanced Research","keywords":"Ambiguity; Initialization; Foundation (evidence); Benchmark (surveying); Classifier (UML); Convolutional neural network; Model selection","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004399809,0.0002374161,0.0002301339,0.0001541558,0.0005096417,0.001012416,0.001786458,0.0002619043,0.00002015078],"category_scores_gemma":[0.000132983,0.0001859118,0.000126973,0.0002882769,0.00008876281,0.001184389,0.001476962,0.0005840206,0.00005016685],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008633892,"about_ca_system_score_gemma":0.0001575863,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002253923,"about_ca_topic_score_gemma":0.000003930313,"domain_scores_codex":[0.9983726,0.0000750332,0.0003529935,0.0006663487,0.0002312997,0.00030179],"domain_scores_gemma":[0.9981626,0.0002225345,0.0003289925,0.001080934,0.0001724434,0.00003248858],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00004003629,0.000172024,0.00893602,0.0002609261,0.0003006918,0.000003084961,0.007975398,0.0130049,0.001082769,0.496939,0.007602456,0.4636827],"study_design_scores_gemma":[0.0004612522,0.00005078742,0.0005827743,0.0002299775,0.00005220599,8.25201e-7,0.000688838,0.826615,0.008868651,0.08995248,0.07206766,0.0004295523],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0371731,0.000455437,0.9335811,0.02514486,0.001459217,0.0006156007,0.000004315885,0.00091465,0.000651782],"genre_scores_gemma":[0.9745687,0.0006185288,0.01487584,0.0009397304,0.0001720202,0.000566361,0.0000833389,0.00001833209,0.008157155],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9373956,"threshold_uncertainty_score":0.976275,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08464897028036827,"score_gpt":0.3101483991326191,"score_spread":0.2254994288522508,"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."}}