{"id":"W2776817984","doi":"10.1109/iesys.2017.8233557","title":"Regularizing the loss layer of CNNs for facial expression recognition using crowdsourced labels","year":2017,"lang":"en","type":"article","venue":"","topic":"Emotion and Mood Recognition","field":"Psychology","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Overfitting; Convolutional neural network; Computer science; Artificial intelligence; Generalization; Pattern recognition (psychology); Layer (electronics); Construct (python library); Ground truth; Machine learning; Deep learning; Training set; Artificial neural network; Mathematics","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":[],"consensus_categories":[],"category_scores_codex":[0.0003359176,0.00009087619,0.0001260785,0.0000469927,0.0005364509,0.00005123912,0.0001657431,0.000138367,0.0009064874],"category_scores_gemma":[0.0001226786,0.00006364415,0.0001009861,0.00002537894,0.0001007557,0.0001330554,0.00004185725,0.00008648413,0.00004574106],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001115696,"about_ca_system_score_gemma":0.00001146545,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008289958,"about_ca_topic_score_gemma":0.00002573838,"domain_scores_codex":[0.999247,0.00007966754,0.0002138514,0.0001874198,0.00009900964,0.0001730157],"domain_scores_gemma":[0.9991113,0.00007593148,0.0002703301,0.0003683101,0.000139947,0.00003416068],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0005095665,0.0003102874,0.001950068,0.00007183979,0.0001249921,0.000003939656,0.003876668,0.000006832372,0.773483,0.001982572,0.006404323,0.2112758],"study_design_scores_gemma":[0.006909649,0.0003150919,0.02342217,0.0004317026,0.0002213576,0.0000535712,0.003414064,0.0008271888,0.9299761,0.01733686,0.01650217,0.0005900476],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9422405,0.00001761038,0.02570507,0.0006980784,0.0009770382,0.0004921858,0.00004866181,0.00004344245,0.02977746],"genre_scores_gemma":[0.9940796,0.000003058301,0.002881268,0.0001871884,0.00026074,0.00003039706,0.00002849967,0.00001741559,0.00251181],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2106858,"threshold_uncertainty_score":0.9925401,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1841134022570582,"score_gpt":0.3935047376218545,"score_spread":0.2093913353647963,"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."}}