{"id":"W4226193037","doi":"10.1609/aaai.v36i11.21452","title":"Evaluating Explainable AI on a Multi-Modal Medical Imaging Task: Can Existing Algorithms Fulfill Clinical Requirements?","year":2022,"lang":"en","type":"article","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia; Simon Fraser University","funders":"BC Cancer Foundation; Compute Canada; Nvidia","keywords":"Computer science; Modality (human–computer interaction); Artificial intelligence; Modal; Leverage (statistics); Machine learning; Metric (unit); Feature (linguistics); Medical imaging; Data mining; Engineering","routes":{"ca_aff":true,"ca_fund":true,"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":["metaepi_narrow","sts","open_science"],"consensus_categories":[],"category_scores_codex":[0.007831675,0.000586075,0.0007132089,0.0004049874,0.001840465,0.0005993588,0.006221225,0.0001575897,0.0005227705],"category_scores_gemma":[0.005897599,0.000516841,0.0003882778,0.001802475,0.0006891139,0.0008568312,0.003063002,0.001922545,0.0001428143],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004547224,"about_ca_system_score_gemma":0.0008587764,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005464047,"about_ca_topic_score_gemma":0.00005792415,"domain_scores_codex":[0.9910172,0.0003432649,0.002228685,0.001737015,0.003366696,0.001307186],"domain_scores_gemma":[0.9951299,0.0007293821,0.00125313,0.0009895228,0.001446308,0.0004517453],"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":[0.0002055124,0.001209668,0.0009013289,0.00005572252,0.0000507319,0.0000359097,0.003657413,0.001590176,0.01168471,0.5876289,0.0006294381,0.3923505],"study_design_scores_gemma":[0.0001334998,0.0008956909,0.00009282611,0.000322622,0.00002519455,0.00004058149,0.0047984,0.836508,0.09405284,0.06195311,0.0005521628,0.0006250899],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.4928995,0.0002646303,0.3765475,0.07834665,0.0120413,0.00599046,0.0001047021,0.00166488,0.03214037],"genre_scores_gemma":[0.9854298,0.00002254178,0.01019334,0.003159474,0.0002862267,0.0002819103,0.00000302151,0.00005422338,0.0005694524],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8349178,"threshold_uncertainty_score":0.9997283,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2772565168370234,"score_gpt":0.4433974170613346,"score_spread":0.1661409002243112,"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."}}