{"id":"W4385572364","doi":"10.18653/v1/2023.findings-acl.590","title":"Zero-shot Visual Question Answering with Language Model Feedback","year":2023,"lang":"en","type":"article","venue":"","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"National Natural Science Foundation of China","keywords":"Closed captioning; Computer science; Leverage (statistics); Language model; Task (project management); Artificial intelligence; Context (archaeology); Question answering; Natural language processing; Code (set theory); Scheme (mathematics); Machine learning; Speech recognition; Image (mathematics); Programming language; 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":[0.000170935,0.00009458762,0.00007726386,0.0001014333,0.0001042425,0.000101211,0.0003714711,0.0000316693,0.00001042939],"category_scores_gemma":[0.00001824316,0.00007743699,0.00001997986,0.0005894365,0.00001837278,0.0002393622,0.0001492126,0.0001235701,0.0004893522],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002348639,"about_ca_system_score_gemma":0.00003147947,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000510218,"about_ca_topic_score_gemma":0.00003424917,"domain_scores_codex":[0.99918,0.00002502337,0.00009760541,0.0003003077,0.0001917558,0.0002053674],"domain_scores_gemma":[0.9994591,0.00004281836,0.00003320712,0.0003625247,0.00003389003,0.00006843651],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001056184,0.00009192865,0.006465614,0.00002844108,0.00002082788,0.00001925826,0.003460993,0.7314746,0.03353409,0.1671676,0.001146138,0.05657995],"study_design_scores_gemma":[0.0001316883,0.00002945984,0.01472154,0.00000914626,0.00000182958,0.000004576351,0.00003045234,0.9834166,0.0009160234,0.000569176,0.00005455634,0.0001149177],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3325238,0.000003121888,0.6610377,0.0008097003,0.00001809696,0.00008438669,5.241133e-7,0.001042907,0.004479755],"genre_scores_gemma":[0.8953331,0.000001350747,0.1026648,0.00014629,0.00001993634,0.00003268811,0.000008529749,0.00001191737,0.001781475],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5628092,"threshold_uncertainty_score":0.6289796,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01670687766372517,"score_gpt":0.3167307179897402,"score_spread":0.300023840326015,"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."}}