{"id":"W4416260299","doi":"10.48550/arxiv.2506.21744","title":"Federated Item Response Models: A Gradient-driven Privacy-preserving Framework for Distributed Psychometric Estimation","year":2025,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Psychometric Methodologies and Testing","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Social Sciences and Humanities Research Council of Canada; Connaught Fund; University of Toronto","keywords":"Item response theory; Estimation; Field (mathematics); Raw data; Information privacy; Data modeling; Federated learning; Privacy protection","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03131553,0.002345018,0.002703657,0.001697857,0.001041217,0.003743869,0.008242554,0.002990107,0.008847506],"category_scores_gemma":[0.1019408,0.001827561,0.002786229,0.003079138,0.002968823,0.006499813,0.008645228,0.005769082,0.004577741],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002435171,"about_ca_system_score_gemma":0.003759908,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003184717,"about_ca_topic_score_gemma":0.003160583,"domain_scores_codex":[0.970749,0.02127507,0.001046009,0.003485842,0.002733602,0.0007104964],"domain_scores_gemma":[0.951762,0.02596228,0.002632684,0.01528656,0.003592133,0.0007643574],"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":[0.0009165678,0.000353363,0.005929311,0.0005235286,0.0004224926,0.0004144988,0.001089085,0.327623,0.002672818,0.3628964,0.01386177,0.2832973],"study_design_scores_gemma":[0.0001043314,0.0001059409,0.0004926887,0.00005512834,0.00004185416,0.000129797,0.00007041566,0.7099754,0.001484584,0.2829468,0.004545233,0.00004790803],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001501669,0.00009327995,0.9963285,0.0003153897,0.00001958145,0.0001254426,0.0003012004,0.0008765211,0.0004383107],"genre_scores_gemma":[0.1634414,0.0003656949,0.8283341,0.0008201022,0.000185764,0.001620276,0.001406578,0.0006085974,0.003217471],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.03131553,"threshold_uncertainty_score":0.1656144,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5639737864270469,"score_gpt":0.3737501621731419,"score_spread":0.1902236242539049,"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."}}