{"id":"W2994853525","doi":"10.48550/arxiv.1912.06960","title":"What Else Can Fool Deep Learning? Addressing Color Constancy Errors on Deep Neural Network Performance","year":2019,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Image Enhancement Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"York University","funders":"Natural Sciences and Engineering Research Council of Canada; Samsung; Canada First Research Excellence Fund; York University","keywords":"Computer science; Artificial intelligence; Color constancy; Robustness (evolution); Deep neural networks; Segmentation; Color balance; Artificial neural network; Deep learning; Computer vision; Image (mathematics); Pattern recognition (psychology); Image processing; Color image","routes":{"ca_aff":true,"ca_fund":true,"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"],"consensus_categories":[],"category_scores_codex":[0.0003895774,0.0005985454,0.0005562419,0.0003222042,0.0004464657,0.0005782275,0.002532627,0.0003898382,0.00004583788],"category_scores_gemma":[0.00004006094,0.0007010282,0.0002306969,0.000777038,0.0002151566,0.001464229,0.002401675,0.001587892,0.0001189799],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005414517,"about_ca_system_score_gemma":0.0001839263,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003562798,"about_ca_topic_score_gemma":0.00008888359,"domain_scores_codex":[0.9966271,0.0003213837,0.0003319975,0.00158687,0.0002424972,0.0008901402],"domain_scores_gemma":[0.9972329,0.0001804472,0.0005739779,0.00158498,0.000233829,0.0001938482],"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":[0.00009109303,0.0000805028,0.006872534,0.0001474029,0.00007833063,0.0003115544,0.000541505,0.9689566,0.0000427305,0.007716062,0.0002467403,0.01491491],"study_design_scores_gemma":[0.0004741197,0.0003193179,0.001423918,0.0006661724,0.00005337456,0.000007569117,0.0001627983,0.993337,0.0007893028,0.001454152,0.000484031,0.0008281988],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5606369,0.0002522411,0.431186,0.0001213843,0.002552472,0.0008201685,0.000002045042,0.0009499327,0.003478897],"genre_scores_gemma":[0.9933059,0.0006126295,0.00329422,0.0004272678,0.0001503917,0.000005742573,0.00002203397,0.0000474046,0.002134418],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.432669,"threshold_uncertainty_score":0.9995441,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05493843091998473,"score_gpt":0.2062150004604581,"score_spread":0.1512765695404734,"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."}}