{"id":"W6966793414","doi":"10.4224/40002946","title":"ImpAcX_OnHW: Improving accuracy and explainability of handwriting recognition models","year":2022,"lang":"en","type":"other","venue":"NPARC","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada; University of Waterloo","funders":"","keywords":"Handwriting; Pattern recognition (psychology); Feature (linguistics); Intelligent character recognition; Handwriting recognition","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"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":[0.002913037,0.002345853,0.0009851065,0.002062953,0.000665257,0.002357354,0.003479358,0.001840182,0.01882385],"category_scores_gemma":[0.02158239,0.0006800979,0.001732865,0.001224778,0.0006820139,0.003174528,0.00277235,0.002888592,0.009213369],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001246367,"about_ca_system_score_gemma":0.001803332,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01421285,"about_ca_topic_score_gemma":0.01366053,"domain_scores_codex":[0.9970779,0.0006841972,0.000149358,0.0008810021,0.001007326,0.0002001308],"domain_scores_gemma":[0.9897251,0.004581771,0.000338256,0.003761272,0.001374668,0.0002189298],"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":[0.001053954,0.0004691781,0.006852295,0.000773236,0.0004466943,0.0003880394,0.0002784891,0.1302303,0.01257745,0.01146655,0.1958153,0.6396484],"study_design_scores_gemma":[0.0001483878,0.0001464451,0.001989027,0.00006521612,0.00007160629,0.0001368121,0.00005073386,0.9326859,0.02612508,0.0132345,0.02529742,0.00004882058],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.090749,0.001892849,0.585642,0.002258566,0.001091458,0.0003506253,0.02478324,0.2772651,0.01596715],"genre_scores_gemma":[0.3683346,0.0009441702,0.4957505,0.0007836013,0.0004019763,0.0005196062,0.08921244,0.02093783,0.02311526],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01882385,"threshold_uncertainty_score":0.06297207,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03397021265383127,"score_gpt":0.2686452068131766,"score_spread":0.2346749941593453,"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."}}