{"id":"W3169447391","doi":"10.48550/arxiv.2106.06620","title":"Robust Representation Learning via Perceptual Similarity Metrics","year":2021,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Generative Adversarial Networks and Image Synthesis","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Overfitting; Computer science; Spurious relationship; Artificial intelligence; Metric (unit); Representation (politics); Machine learning; Feature learning; Similarity (geometry); Task (project management); Information bottleneck method; Bottleneck; Feature (linguistics); Generalization; Transformation (genetics); Mutual information; Pattern recognition (psychology); 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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.000372151,0.0003326779,0.0004138275,0.000293943,0.0003515618,0.0004178666,0.001252997,0.0003350684,0.0001280803],"category_scores_gemma":[0.0002481813,0.0004046728,0.0003289289,0.001393711,0.00009777625,0.0007712926,0.002574973,0.0009706158,0.00003595673],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002214492,"about_ca_system_score_gemma":0.0001814088,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004144002,"about_ca_topic_score_gemma":0.00007963502,"domain_scores_codex":[0.997125,0.0006298302,0.0002367932,0.001458957,0.0001709935,0.0003784731],"domain_scores_gemma":[0.9978537,0.0002339158,0.0002786614,0.001048729,0.0004109631,0.0001740532],"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":[0.000007578435,0.00006395923,0.002726399,0.0000184496,0.00009237923,0.0002446928,0.0004412862,0.9916593,0.0001220427,0.001550704,0.0002346302,0.00283854],"study_design_scores_gemma":[0.0002261144,0.00003329036,0.002271138,0.0000320988,0.00008212609,0.000004258735,0.0004491561,0.9943551,0.0004362942,0.001420917,0.0002609859,0.0004284862],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05285225,0.000104363,0.9440341,0.0001134617,0.0008213105,0.0001759151,0.000003030005,0.0001923697,0.001703187],"genre_scores_gemma":[0.9765259,0.0002957422,0.02156337,0.00008858653,0.0001929807,7.960712e-7,0.00004933618,0.00001916244,0.001264123],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9236736,"threshold_uncertainty_score":0.9998405,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1312729489414823,"score_gpt":0.2003486580711475,"score_spread":0.06907570912966518,"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."}}