{"id":"W4403487119","doi":"10.3233/faia240840","title":"Reset It and Forget It: Relearning Last-Layer Weights Improves Continual and Transfer Learning","year":2024,"lang":"en","type":"book-chapter","venue":"Frontiers in artificial intelligence and applications","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vector Institute; University of British Columbia; Canadian Institute for Advanced Research","funders":"","keywords":"Reset (finance); Layer (electronics); Computer science; Business; Materials science; Nanotechnology","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":[0.0006521959,0.0007527336,0.0005775696,0.0003567594,0.0002749995,0.0009792476,0.001698352,0.000808689,0.004631032],"category_scores_gemma":[0.003296777,0.0003223932,0.0004835242,0.0004092527,0.0007386129,0.002870132,0.001516359,0.001914658,0.001411994],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004577689,"about_ca_system_score_gemma":0.0004089923,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001460583,"about_ca_topic_score_gemma":0.001823605,"domain_scores_codex":[0.9997727,0.00004264146,0.00001115084,0.00009095491,0.00005361721,0.00002881017],"domain_scores_gemma":[0.9991919,0.0002960363,0.00005657738,0.0003068337,0.0000909365,0.00005777113],"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":[0.0001521699,0.0001925489,0.001478373,0.0001626309,0.00007810775,0.0001195893,0.0002668816,0.09616663,0.02868928,0.01606809,0.006166341,0.8504595],"study_design_scores_gemma":[0.00001835989,0.0002148568,0.0008428165,0.00004915999,0.00005051312,0.0001997075,0.00006116491,0.9432183,0.01768355,0.03139293,0.006239644,0.00002910473],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1123388,0.001742212,0.8665035,0.0004527506,0.0002860818,0.00007992383,0.0001424386,0.00624377,0.01221052],"genre_scores_gemma":[0.7335677,0.000601185,0.2532536,0.0003354181,0.00008789019,0.00007495297,0.0003808125,0.0005062695,0.01119226],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004631032,"threshold_uncertainty_score":0.01549238,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0365648664025473,"score_gpt":0.2768199104064673,"score_spread":0.2402550440039199,"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."}}