{"id":"W4403990800","doi":"10.1007/978-3-031-72986-7_25","title":"TIBET: Identifying and Evaluating Biases in Text-to-Image Generative Models","year":2024,"lang":"en","type":"book-chapter","venue":"Lecture notes in computer science","topic":"Multimodal Machine Learning Applications","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":false,"ca_institutions":"Vector Institute; University of British Columbia","funders":"","keywords":"Computer science; Generative grammar; Image (mathematics); Artificial intelligence; Generative model; Natural language processing; Information retrieval","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":[0.009232542,0.001692334,0.001577553,0.001893862,0.0009228851,0.002906438,0.002899178,0.002794757,0.005619152],"category_scores_gemma":[0.04219297,0.0009311536,0.001808019,0.001774871,0.001395729,0.003929481,0.00336143,0.003908893,0.00207563],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001874231,"about_ca_system_score_gemma":0.001540997,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008169123,"about_ca_topic_score_gemma":0.01105369,"domain_scores_codex":[0.9965848,0.002109976,0.00013659,0.0004849465,0.0005227335,0.0001609275],"domain_scores_gemma":[0.9648563,0.0310852,0.0005619113,0.001968032,0.001161563,0.0003670809],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"observational","study_design_scores_codex":[0.00204642,0.0004315964,0.01070423,0.0006910111,0.0007673355,0.0002345591,0.0004731397,0.4946997,0.00428992,0.03600494,0.03246202,0.4171952],"study_design_scores_gemma":[0.0000598308,0.00005293315,0.0003487042,0.00002834888,0.0000390277,0.00003157246,0.00002677772,0.9706454,0.00134452,0.02649659,0.0009138574,0.00001237616],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1251292,0.002646782,0.8421901,0.0009607931,0.000413886,0.0003661617,0.003002804,0.0190996,0.00619074],"genre_scores_gemma":[0.6334155,0.0007316528,0.3450535,0.0007813476,0.0002736218,0.0005308156,0.009638886,0.003659908,0.005914828],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009232542,"threshold_uncertainty_score":0.04882687,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0669570626640727,"score_gpt":0.353602100343215,"score_spread":0.2866450376791423,"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."}}