{"id":"W4387618741","doi":"10.1016/j.patter.2023.100844","title":"Nonlinear independent component analysis for principled disentanglement in unsupervised deep learning","year":2023,"lang":"en","type":"review","venue":"Patterns","topic":"Blind Source Separation Techniques","field":"Computer Science","cited_by":28,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Precursory Research for Embryonic Science and Technology; Japan Science and Technology Agency; Japan Society for the Promotion of Science; Academy of Finland; Canadian Institute for Advanced Research","keywords":"Independent component analysis; Identifiability; Nonlinear system; Heuristic; Artificial intelligence; Computer science; Unsupervised learning; Representation (politics); Component (thermodynamics); Feature learning; Machine learning; Probabilistic logic; Uniqueness; Deep learning; Pattern recognition (psychology); Algorithm; Mathematics","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001033411,0.0004068461,0.001432866,0.001228594,0.00008608623,0.00023904,0.001287068,0.000203195,0.00002607186],"category_scores_gemma":[0.00004517857,0.0003704434,0.0008035246,0.001475451,0.00001157736,0.0001426758,0.0006525369,0.0004795403,0.00006051049],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000299589,"about_ca_system_score_gemma":0.00009982129,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008781387,"about_ca_topic_score_gemma":0.0007368534,"domain_scores_codex":[0.996797,0.0004337964,0.001015604,0.0008330279,0.0004952666,0.0004252524],"domain_scores_gemma":[0.9982808,0.0003386906,0.00044826,0.0007523853,0.00006832467,0.0001114867],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000005344134,0.0004524073,0.00600409,0.009285972,0.001836562,0.00007801344,0.002156892,0.001111656,9.187474e-7,0.0016452,0.00002393161,0.977399],"study_design_scores_gemma":[0.000742041,0.0001696251,0.00129002,0.003738473,0.001358296,0.000005399791,0.00008583808,0.1733231,0.00001382374,0.0001277473,0.8179016,0.001244024],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"review","genre_scores_codex":[0.0001125645,0.226976,0.7700147,0.00008835264,0.000150358,0.002006454,0.00006299552,0.0005524289,0.00003618071],"genre_scores_gemma":[0.000492006,0.9870619,0.009588607,0.00009073352,0.00009346557,0.001344065,0.0010423,0.0000737327,0.0002131877],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.976155,"threshold_uncertainty_score":0.9998748,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09468974548462399,"score_gpt":0.374004057197054,"score_spread":0.27931431171243,"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."}}