{"id":"W3015943301","doi":"10.1109/ijcnn48605.2020.9206833","title":"Fisher Discriminant Triplet and Contrastive Losses for Training Siamese Networks","year":2020,"lang":"en","type":"article","venue":"","topic":"AI in cancer detection","field":"Computer Science","cited_by":35,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Linear discriminant analysis; MNIST database; Pattern recognition (psychology); Discriminant; Artificial intelligence; Subspace topology; Computer science; Metric (unit); Artificial neural network; Feature (linguistics); Kernel Fisher discriminant analysis; Multiple discriminant analysis; Feature extraction; Fisher kernel; Mathematics; Engineering","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":[0.002855851,0.001432937,0.0008535095,0.001096933,0.0004046471,0.0007651989,0.00149564,0.001386093,0.003151826],"category_scores_gemma":[0.007273987,0.0003085322,0.0006647764,0.0008750733,0.0009459443,0.001758412,0.001298756,0.00252749,0.0008881201],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001291326,"about_ca_system_score_gemma":0.001213749,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004729662,"about_ca_topic_score_gemma":0.005204734,"domain_scores_codex":[0.9991316,0.0003066616,0.00005476893,0.0001824421,0.000241084,0.00008349043],"domain_scores_gemma":[0.9983522,0.0008116556,0.0001126511,0.0002197268,0.0004023502,0.000101375],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002797211,0.0002869461,0.002108345,0.00009461251,0.00009047658,0.0001051191,0.0000543458,0.6328198,0.007894888,0.02420973,0.008716466,0.3233395],"study_design_scores_gemma":[0.000006311014,0.00003051561,0.0001355651,0.000003774478,0.000003050535,0.0000173181,0.000003839061,0.9945096,0.001009923,0.003835606,0.0004390736,0.00000553096],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02676402,0.0004367304,0.9694368,0.0003665441,0.00006552414,0.00007705603,0.0001612019,0.0008005964,0.001891616],"genre_scores_gemma":[0.515473,0.0004526967,0.4726476,0.0005357256,0.0001212443,0.0004564673,0.001288677,0.0002985577,0.008725959],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004729662,"threshold_uncertainty_score":0.0151034,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05824717489090631,"score_gpt":0.2528135293446994,"score_spread":0.1945663544537931,"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."}}