{"id":"W4415482185","doi":"10.1109/tsp.2025.3624791","title":"Regularized Top-$ k $: A Bayesian Framework for Gradient Sparsification","year":2025,"lang":"","type":"article","venue":"IEEE Transactions on Signal Processing","topic":"Medical Image Segmentation Techniques","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Ericsson (Canada); University of Toronto","funders":"","keywords":"Scaling; Posterior probability; Convergence (economics); Rate of convergence; Bayesian probability; Prior probability; Generalization; Inverse; Inverse problem","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":[0.0008872173,0.0004643596,0.0004916725,0.0007283028,0.001253732,0.0008907198,0.0009741602,0.000434452,0.0001945129],"category_scores_gemma":[0.00005230102,0.0005018995,0.0003274259,0.001946712,0.0003171769,0.0009692911,0.000005921549,0.0008105178,0.00001995386],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003650851,"about_ca_system_score_gemma":0.0007651521,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001203377,"about_ca_topic_score_gemma":0.000003153972,"domain_scores_codex":[0.9961289,0.0002283131,0.00104673,0.001213228,0.0007026345,0.000680159],"domain_scores_gemma":[0.9975219,0.0005627932,0.0004035555,0.0007208488,0.0005117601,0.0002791579],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002073297,0.0006063894,0.000001224916,0.0005328947,0.00007630356,0.000003866636,0.001258038,0.0009091193,0.008170728,0.002993508,0.0002839974,0.9849566],"study_design_scores_gemma":[0.0008876721,0.0003139483,0.000008677928,0.002011759,0.0001808636,0.000006035114,0.0001823622,0.4853877,0.4719369,0.03835602,0.0003038169,0.0004241925],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.00006782868,0.00034509,0.9923266,0.003789324,0.001079402,0.001606859,0.00002035958,0.0004824981,0.000281997],"genre_scores_gemma":[0.5050623,0.00005014124,0.4918438,0.001542814,0.00007695707,0.000431728,0.000003475814,0.00002740284,0.0009614051],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9845324,"threshold_uncertainty_score":0.9997433,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02870761014117459,"score_gpt":0.3225312797731015,"score_spread":0.2938236696319269,"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."}}