{"id":"W4411399751","doi":"10.1016/j.patcog.2025.111975","title":"Refining pseudo-labels through iterative mix-up for weakly supervised semantic segmentation","year":2025,"lang":"en","type":"article","venue":"Pattern Recognition","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"Training Program for Excellent Young Innovators of Changsha; Higher Education Discipline Innovation Project; Natural Science Foundation of Hunan Province; National Science Foundation","keywords":"Refining (metallurgy); Segmentation; Artificial intelligence; Computer science; Pattern recognition (psychology); Natural language processing; Computer vision; Materials 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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001237775,0.0001670974,0.000160682,0.00009521753,0.0002774809,0.0001439879,0.0003004813,0.00006147939,0.000017582],"category_scores_gemma":[0.00003677749,0.0001752471,0.00006817695,0.0004687941,0.00002587542,0.0008712492,0.00009097637,0.0001111775,0.00009158906],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007226328,"about_ca_system_score_gemma":0.00002876329,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001547852,"about_ca_topic_score_gemma":0.00002683894,"domain_scores_codex":[0.9986953,0.00007426723,0.0003147292,0.0005066217,0.0001375726,0.0002715569],"domain_scores_gemma":[0.9990035,0.000328384,0.0001306531,0.0003109935,0.0001908433,0.00003565753],"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":[0.00001803076,0.00007556161,0.0005485427,0.00009382523,0.00003945738,0.000001628563,0.001043169,0.00008316644,0.02695397,0.001489908,0.001452051,0.9682007],"study_design_scores_gemma":[0.00549497,0.0004865914,0.004131691,0.001197829,0.000192492,0.00003102503,0.0005068163,0.3973588,0.3668959,0.2175866,0.004648601,0.001468662],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04643613,0.00005085233,0.9494824,0.002126743,0.0005064051,0.0005792548,0.00003698553,0.0002118604,0.0005693177],"genre_scores_gemma":[0.7322147,0.00006568358,0.2599015,0.004873434,0.0001767015,0.001512444,0.0004974512,0.00002804473,0.0007300817],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.966732,"threshold_uncertainty_score":0.7146368,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05057411921813759,"score_gpt":0.3192396139583518,"score_spread":0.2686654947402142,"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."}}