{"id":"W4410426493","doi":"10.2139/ssrn.5257431","title":"Improving Vision Transformers by Overlapping Heads in Multi-Head Self-Attention","year":2025,"lang":"en","type":"preprint","venue":"SSRN Electronic Journal","topic":"Brain Tumor Detection and Classification","field":"Neuroscience","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"Toronto Metropolitan University","funders":"","keywords":"Transformer; Head (geology); Computer vision; Computer science; Artificial intelligence; Optometry; Engineering; Electrical engineering; Medicine; Geology; Voltage","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":["metaepi_narrow","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.001800931,0.0003432559,0.0003346783,0.000522511,0.000344124,0.000210692,0.0004592226,0.0003587765,0.00001426695],"category_scores_gemma":[0.0001754608,0.0003568348,0.0002644409,0.0004608647,0.00004701298,0.0003268653,0.00009355588,0.006235899,0.00001526269],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003585825,"about_ca_system_score_gemma":0.002388602,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001739792,"about_ca_topic_score_gemma":0.0009528362,"domain_scores_codex":[0.9958152,0.0003957302,0.0007059558,0.0007637333,0.0004551202,0.001864269],"domain_scores_gemma":[0.9990637,0.00008519406,0.000446581,0.0002499718,0.00005899294,0.00009559054],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001500539,0.000561968,0.0005181825,0.0002077137,0.00004738792,0.000006174266,0.000383537,0.000237374,0.7629486,0.004872659,0.0000737957,0.2299926],"study_design_scores_gemma":[0.02859629,0.003117218,0.02345092,0.004793712,0.0007004512,0.002699404,0.01088503,0.605826,0.1866452,0.1072622,0.01874073,0.007282835],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6610987,0.00227918,0.3276069,0.003098442,0.002730326,0.001310564,0.00003817324,0.0004109063,0.001426872],"genre_scores_gemma":[0.9914066,0.00610086,0.0001424827,0.0002513575,0.0000907549,0.00004137149,0.0000124124,0.00003226219,0.001921895],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6055886,"threshold_uncertainty_score":0.9998884,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01904065188785814,"score_gpt":0.2948945107666868,"score_spread":0.2758538588788287,"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."}}