{"id":"W3202332982","doi":"10.1609/aaai.v36i10.21306","title":"Retrieve, Caption, Generate: Visual Grounding for Enhancing Commonsense in Text Generation Models","year":2022,"lang":"en","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Closed captioning; Commonsense reasoning; Computer science; Fluency; Transformer; Generative grammar; Commonsense knowledge; Artificial intelligence; Natural language processing; Image (mathematics); Linguistics; Knowledge extraction; 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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001346571,0.001166284,0.0003997745,0.0009052554,0.0003321483,0.001384437,0.001508469,0.001094448,0.0100532],"category_scores_gemma":[0.007408188,0.0003445507,0.0007654217,0.0004119351,0.0007566134,0.002973662,0.001691809,0.001388264,0.002075275],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006270796,"about_ca_system_score_gemma":0.0004866015,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001975688,"about_ca_topic_score_gemma":0.003319059,"domain_scores_codex":[0.9995271,0.0001828818,0.000021463,0.0001349821,0.00009863231,0.00003491486],"domain_scores_gemma":[0.997759,0.001406046,0.0001001786,0.0004854332,0.000175208,0.00007418661],"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":[0.000678947,0.0005634501,0.002561894,0.0007577196,0.0001181907,0.0006854215,0.001521691,0.2048738,0.05373548,0.04232254,0.02347974,0.6687012],"study_design_scores_gemma":[0.00007463121,0.0001740408,0.0004053191,0.00004317791,0.00004649406,0.000212334,0.0001516532,0.9127682,0.04306646,0.02897472,0.01404731,0.0000357411],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06916484,0.0004186239,0.8836517,0.0007041393,0.0001773987,0.0004157307,0.001436961,0.03143993,0.01259065],"genre_scores_gemma":[0.5603633,0.0001975308,0.4295201,0.0003587585,0.00005712394,0.0002649884,0.002856429,0.0018587,0.004523066],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.0100532,"threshold_uncertainty_score":0.03363138,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04270963079109104,"score_gpt":0.3035675939309457,"score_spread":0.2608579631398547,"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."}}