{"id":"W4385261898","doi":"10.48550/arxiv.2307.12698","title":"MC-JEPA: A Joint-Embedding Predictive Architecture for Self-Supervised Learning of Motion and Content Features","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Advanced Vision and Imaging","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Agence Nationale de la Recherche; Division of Mathematical Sciences; York University","keywords":"Computer science; Optical flow; Artificial intelligence; Embedding; Supervised learning; Encoder; Segmentation; Machine learning; Unsupervised learning; Motion (physics); Multimodal learning; Feature learning; Object (grammar); Joint (building); Focus (optics); Pattern recognition (psychology); Semi-supervised learning; Image (mathematics); Artificial neural network","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":[0.0002381717,0.0002550806,0.0003660994,0.0003447707,0.0001885398,0.00007149587,0.0005319338,0.0001595299,0.000001897157],"category_scores_gemma":[0.0001451399,0.0002688466,0.0001879609,0.000308982,0.00007445199,0.0002842748,0.001292524,0.0005924089,0.00000222581],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009584735,"about_ca_system_score_gemma":0.00005601811,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003624289,"about_ca_topic_score_gemma":0.000005332888,"domain_scores_codex":[0.9984288,0.0001096874,0.0001947327,0.0008916616,0.00009662304,0.000278468],"domain_scores_gemma":[0.9987627,0.0001927577,0.000274815,0.0004316189,0.000222181,0.0001158808],"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":[0.0001780008,0.0001325376,0.003713484,0.0008394571,0.0003663281,0.000102234,0.005073365,0.9163913,0.002321221,0.03764291,0.0002157558,0.03302337],"study_design_scores_gemma":[0.0007358472,0.0001184931,0.002683522,0.0002588979,0.0000582475,0.00000536737,0.0005439437,0.9812742,0.0005045272,0.01346972,0.00009235605,0.0002548836],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05109548,0.00008508941,0.9472805,0.0001639941,0.0002957787,0.0005110669,0.00001752431,0.0004093279,0.0001412062],"genre_scores_gemma":[0.9524585,0.0001578004,0.04664953,0.00004363524,0.00004841514,0.000003165985,0.00001604505,0.00002432573,0.000598572],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.901363,"threshold_uncertainty_score":0.9999764,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09058569629909079,"score_gpt":0.2210678516634994,"score_spread":0.1304821553644086,"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."}}