{"id":"W4413145840","doi":"10.1109/cvpr52734.2025.00877","title":"DivPrune: Diversity-based Visual Token Pruning for Large Multimodal Models","year":2025,"lang":"en","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"Huawei Technologies (Canada)","funders":"","keywords":"Computer science; Security token; Pruning; Diversity (politics); Artificial intelligence; Computer network; Biology","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":[0.001595686,0.001978164,0.001693859,0.00108712,0.0008849291,0.001543504,0.003578133,0.00176446,0.005039623],"category_scores_gemma":[0.005676356,0.0009969035,0.001674855,0.0008790385,0.0009673568,0.002489092,0.00244938,0.002442789,0.001638151],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001487329,"about_ca_system_score_gemma":0.001895695,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01048226,"about_ca_topic_score_gemma":0.01939018,"domain_scores_codex":[0.9991949,0.0002028034,0.00005204489,0.0002411816,0.0001870239,0.0001220504],"domain_scores_gemma":[0.9981253,0.001104839,0.0001435217,0.0003076906,0.0002163849,0.0001022957],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0006002898,0.0001955984,0.002575644,0.0004135726,0.0002390477,0.000626235,0.0002921099,0.5321742,0.01231221,0.01132793,0.01978629,0.4194569],"study_design_scores_gemma":[0.00002337101,0.0000333211,0.0001085075,0.00001467822,0.00001336351,0.00008636639,0.00002499892,0.9886491,0.003525912,0.005970384,0.001540434,0.000009469155],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0259463,0.0008655465,0.9595418,0.0003324185,0.00008515186,0.0001758895,0.000723884,0.01052797,0.00180115],"genre_scores_gemma":[0.3166347,0.0003721629,0.671697,0.0004336646,0.00008541447,0.0003834372,0.003447379,0.00189193,0.005054242],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01048226,"threshold_uncertainty_score":0.02084249,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01616403560729983,"score_gpt":0.3084121409683172,"score_spread":0.2922481053610174,"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."}}