{"id":"W4283805152","doi":"10.1609/aaai.v36i2.20123","title":"Improving Zero-Shot Phrase Grounding via Reasoning on External Knowledge and Spatial Relations","year":2022,"lang":"en","type":"article","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"Queen's University","funders":"","keywords":"Phrase; Computer science; Noun phrase; Artificial intelligence; Ground; Visual reasoning; Natural language processing; Modal; Noun; Engineering","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":[0.0007469114,0.0002141497,0.0001993672,0.0002005888,0.001103275,0.0002610757,0.001500604,0.00004920814,0.00009162783],"category_scores_gemma":[0.0004352316,0.0001919638,0.00008058464,0.0006160505,0.0001522015,0.0002952781,0.0009919889,0.0007432819,0.00005007889],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001318094,"about_ca_system_score_gemma":0.00008996032,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008130922,"about_ca_topic_score_gemma":0.00001871245,"domain_scores_codex":[0.9980992,0.00005330259,0.0004471016,0.0006161156,0.0004694069,0.0003148685],"domain_scores_gemma":[0.9987004,0.0001833568,0.0004530728,0.0003326148,0.0002258326,0.0001047474],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00003891971,0.0001602186,0.001567776,0.00001766673,0.00000723605,4.451885e-7,0.001977759,0.0005403792,0.05395478,0.7174312,0.0000192341,0.2242844],"study_design_scores_gemma":[0.00005488781,0.0002920445,0.004257427,0.0001222495,0.00001404658,0.00001933229,0.0002503716,0.8827375,0.0340874,0.07776742,0.0001154007,0.000281887],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4876857,0.0000387726,0.5011719,0.002037569,0.0004504991,0.0005584654,0.000007015562,0.0001793838,0.007870695],"genre_scores_gemma":[0.9925399,0.000003924237,0.006913681,0.000106153,0.00007102425,0.0001134736,7.721232e-7,0.0000175856,0.0002334877],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8821971,"threshold_uncertainty_score":0.8485611,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04921381018470201,"score_gpt":0.305775941823595,"score_spread":0.256562131638893,"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."}}