{"id":"W4416384622","doi":"10.1016/j.engappai.2025.113111","title":"Trustworthy requirements for foundation models—A comprehensive survey and roadmap","year":2025,"lang":"en","type":"article","venue":"Engineering Applications of Artificial Intelligence","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Carleton University","funders":"Research Ireland; Canada Research Chairs; Science Foundation Ireland","keywords":"Trustworthiness; Foundation (evidence); Process (computing); Adversarial system; Generalization","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":[],"consensus_categories":[],"category_scores_codex":[0.0003363685,0.0001112329,0.0001539577,0.000182948,0.0001251118,0.0000640065,0.0004655848,0.00004893307,0.000002091442],"category_scores_gemma":[0.0001209833,0.0001304043,0.0000332898,0.0005889652,0.00004562188,0.0002700129,0.0001533983,0.00009531425,0.000004050066],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003397401,"about_ca_system_score_gemma":0.00004170982,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006932708,"about_ca_topic_score_gemma":0.00001155298,"domain_scores_codex":[0.9990541,0.00002528173,0.0003481902,0.0003073369,0.0001118911,0.0001532198],"domain_scores_gemma":[0.9987782,0.0004520849,0.00009822536,0.0003857742,0.0002512997,0.00003437754],"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":[0.000006396168,0.00001997761,0.00008724158,0.00003428261,0.00001455796,3.841614e-8,0.00007886485,0.47493,0.0005629574,0.4105873,0.00001064521,0.1136677],"study_design_scores_gemma":[0.00002568542,0.00001729771,0.0005912145,0.00001934864,0.000006010186,2.928413e-7,0.00002638043,0.9591301,0.002938947,0.03630754,0.0008359239,0.0001012213],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003109659,0.0001051196,0.9957926,0.000146545,0.0001657035,0.0004833791,0.000004155436,0.0001001772,0.00009269638],"genre_scores_gemma":[0.7894056,0.00001570649,0.210313,0.00002002323,0.00002579142,0.0001630841,0.00001411157,0.000007488325,0.00003512312],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.786296,"threshold_uncertainty_score":0.5317732,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05329427709778476,"score_gpt":0.3276717843211859,"score_spread":0.2743775072234011,"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."}}