{"id":"W2794772564","doi":"10.48550/arxiv.1803.09569","title":"On Regularized Losses for Weakly-supervised CNN Segmentation","year":2018,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"ca_institutions":"École de Technologie Supérieure; University of Waterloo","funders":"","keywords":"Regularization (linguistics); Segmentation; Inference; Computer science; CRFS; Artificial intelligence; Minification; Graph; Machine learning; Pattern recognition (psychology); Algorithm; Conditional random field; Theoretical computer science","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003625337,0.002157119,0.001340926,0.0009429524,0.0006068098,0.001885881,0.002803758,0.002516781,0.00340416],"category_scores_gemma":[0.0125245,0.0008658263,0.0008898786,0.0006688308,0.00212655,0.003596165,0.003271881,0.003068865,0.001597265],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002172875,"about_ca_system_score_gemma":0.001606198,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003023528,"about_ca_topic_score_gemma":0.003381394,"domain_scores_codex":[0.99808,0.0006526963,0.00008496855,0.0005165644,0.0004702819,0.0001954067],"domain_scores_gemma":[0.9964787,0.001838005,0.0002924487,0.0008250315,0.0004028519,0.0001629434],"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":[0.0005170882,0.0001363457,0.00111005,0.00016948,0.00007896712,0.0001408793,0.0001661855,0.8019225,0.01074025,0.03246041,0.005415527,0.1471423],"study_design_scores_gemma":[0.00001132961,0.00003698497,0.0001005938,0.00001338692,0.000006413363,0.00004066876,0.000007653921,0.9805423,0.002780259,0.01586979,0.0005831582,0.000007571505],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02982563,0.0003631031,0.9634757,0.0003585681,0.00004378538,0.00007352263,0.0001593071,0.003038357,0.002662119],"genre_scores_gemma":[0.5777367,0.0004003519,0.4096371,0.0006614447,0.0001186413,0.0002814266,0.001342733,0.001623942,0.008197631],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003625337,"threshold_uncertainty_score":0.01917285,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07747897837039533,"score_gpt":0.2227127744818481,"score_spread":0.1452337961114528,"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."}}