{"id":"W4413145364","doi":"10.1109/cvpr52734.2025.00360","title":"VISTA: Enhancing Long-Duration and High-Resolution Video Understanding by VIdeo SpatioTemporal Augmentation","year":2025,"lang":"en","type":"article","venue":"","topic":"Advanced Image Processing Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Duration (music); 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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003111267,0.0001311923,0.0001183736,0.0001701451,0.0002872481,0.0003876677,0.0002041413,0.00006051137,0.000004879669],"category_scores_gemma":[0.00007609715,0.0001345963,0.00001728447,0.0004120963,0.00005472642,0.002064251,0.0001520535,0.00009381781,0.000002501022],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004050662,"about_ca_system_score_gemma":0.00006499163,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001068205,"about_ca_topic_score_gemma":0.00009341012,"domain_scores_codex":[0.9988534,0.00004883767,0.0003259207,0.0003998619,0.0001867241,0.0001851971],"domain_scores_gemma":[0.9994348,0.00008031847,0.0001493387,0.000224861,0.00007205753,0.00003856717],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00005675364,0.0001118725,0.002070674,0.0002745659,0.00004436263,0.00000812837,0.0009028795,0.00008798407,0.3847373,0.5134322,0.02069671,0.0775766],"study_design_scores_gemma":[0.0007295971,0.000116093,0.001082388,0.0002629943,0.00001912494,0.000006152376,0.0001809328,0.3206295,0.4324377,0.2436873,0.0004341914,0.0004140832],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.001632152,0.000221363,0.9941722,0.002396619,0.0001381034,0.0001975537,9.869185e-7,0.0005994973,0.0006415251],"genre_scores_gemma":[0.63227,0.00002801508,0.3670077,0.0002950161,0.00001155759,0.00001530834,0.00001142284,0.000005204962,0.0003557102],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6306379,"threshold_uncertainty_score":0.5488676,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01602883848453157,"score_gpt":0.2835980329974205,"score_spread":0.2675691945128889,"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."}}