{"id":"W4411119193","doi":"10.18653/v1/2025.naacl-srw.37","title":"ELIOT: Zero-Shot Video-Text Retrieval through Relevance-Boosted Captioning and Structural Information Extraction","year":2025,"lang":"en","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Closed captioning; Shot (pellet); Relevance (law); Computer science; Zero (linguistics); Information retrieval; Feature extraction; Artificial intelligence; Extraction (chemistry); Natural language processing; Speech recognition; Image (mathematics); Linguistics","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":[0.001188463,0.002912283,0.002377494,0.006353855,0.00109506,0.002287846,0.003287027,0.002363036,0.01738355],"category_scores_gemma":[0.002937487,0.0006240065,0.001340562,0.003617419,0.0007266554,0.00388553,0.002519591,0.001540175,0.01775123],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007790556,"about_ca_system_score_gemma":0.001286841,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006651539,"about_ca_topic_score_gemma":0.01110371,"domain_scores_codex":[0.9988499,0.0001877837,0.00008049492,0.0003320126,0.00038052,0.0001692714],"domain_scores_gemma":[0.998566,0.0003874567,0.00009975968,0.0003443672,0.0004923529,0.0001101784],"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":[0.0008604269,0.000353106,0.0003197464,0.0007176812,0.0001826452,0.0003887397,0.0001216395,0.002628991,0.08146675,0.002007726,0.1855005,0.7254521],"study_design_scores_gemma":[0.0007215926,0.001154471,0.003425742,0.0001719142,0.0005367778,0.001615433,0.0005436706,0.660489,0.1946054,0.01449637,0.1219181,0.0003215111],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02935508,0.01350659,0.806864,0.001214666,0.003731098,0.001163207,0.01763874,0.1117584,0.01476832],"genre_scores_gemma":[0.1031275,0.003296688,0.7858784,0.0008694918,0.001463211,0.0007757827,0.06431156,0.003630565,0.03664681],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01738355,"threshold_uncertainty_score":0.05815375,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01266419305219626,"score_gpt":0.3036114091008882,"score_spread":0.290947216048692,"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."}}