{"id":"W4413407284","doi":"10.1016/j.neunet.2025.108017","title":"Rethinking softmax in incremental learning","year":2025,"lang":"en","type":"article","venue":"Neural Networks","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Toronto","funders":"Alliance de recherche numérique du Canada; Natural Sciences and Engineering Research Council of Canada; National Natural Science Foundation of China","keywords":"Softmax function; Artificial intelligence; Computer science; Incremental learning; Machine learning; Pattern recognition (psychology); Artificial neural network","routes":{"ca_aff":true,"ca_fund":true,"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":[0.002837794,0.0007609392,0.001375166,0.0006540944,0.000521219,0.001174821,0.002900667,0.001498602,0.002939798],"category_scores_gemma":[0.0126166,0.0006963342,0.0006580009,0.0006564548,0.001019434,0.003808094,0.002473074,0.003183205,0.0009064297],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007018162,"about_ca_system_score_gemma":0.001074168,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004917386,"about_ca_topic_score_gemma":0.008131136,"domain_scores_codex":[0.9989333,0.0004129446,0.00006359604,0.0002995292,0.0002181417,0.00007249582],"domain_scores_gemma":[0.994749,0.003757162,0.0001180478,0.0007357142,0.0005294163,0.0001105955],"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":[0.0005406188,0.0003595269,0.002080494,0.0002717945,0.0002638587,0.0001272704,0.000228031,0.2945996,0.01059738,0.03052028,0.007218295,0.6531927],"study_design_scores_gemma":[0.000008626929,0.00002795141,0.0001957527,0.00001334612,0.00001702724,0.00001932939,0.00001078272,0.977722,0.002019244,0.01916606,0.000790432,0.000009334394],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01867805,0.001273158,0.9764079,0.0004839442,0.0001904071,0.00003784574,0.0000835851,0.001625668,0.001219473],"genre_scores_gemma":[0.597209,0.0008454604,0.393207,0.00083426,0.0002531499,0.000151659,0.0004944314,0.0006008858,0.006404154],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004917386,"threshold_uncertainty_score":0.01500785,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01647509025716623,"score_gpt":0.2534884894029195,"score_spread":0.2370133991457532,"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."}}