{"id":"W4414359739","doi":"10.24963/ijcai.2025/587","title":"Towards Debiased Generalized Category Discovery","year":2025,"lang":"en","type":"article","venue":"","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Western University","funders":"","keywords":"Overconfidence effect; Classifier (UML); Pattern recognition (psychology); Robustness (evolution); Synthetic data; Training set","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":[0.007385302,0.001499512,0.002799139,0.003369002,0.001643424,0.002458749,0.004143181,0.00270805,0.001508864],"category_scores_gemma":[0.01797887,0.0006544377,0.001280957,0.003050213,0.002305415,0.004750923,0.005473444,0.003429017,0.000920005],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001667092,"about_ca_system_score_gemma":0.003153859,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005225011,"about_ca_topic_score_gemma":0.006004788,"domain_scores_codex":[0.9947173,0.001489424,0.0002538806,0.001650451,0.001481151,0.0004077805],"domain_scores_gemma":[0.9878212,0.004784283,0.001012079,0.003225862,0.002654701,0.0005018708],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0007070664,0.0002519934,0.01130787,0.0002444436,0.0002072976,0.0002252263,0.0005397904,0.07361982,0.008161073,0.01838088,0.009738909,0.8766156],"study_design_scores_gemma":[0.00006035776,0.000189205,0.001464781,0.00004927539,0.0000614297,0.0003545143,0.0002442792,0.9225914,0.008189302,0.0607421,0.006002611,0.00005070699],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06971798,0.001445042,0.9234129,0.0007794049,0.0000944317,0.0001122437,0.0002163192,0.002464087,0.001757632],"genre_scores_gemma":[0.5528829,0.0004537248,0.4392419,0.001232345,0.0001961773,0.0001679365,0.001461646,0.0003373649,0.00402591],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007385302,"threshold_uncertainty_score":0.03905767,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01947617631959016,"score_gpt":0.2708247524718138,"score_spread":0.2513485761522237,"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."}}