{"id":"W4320923498","doi":"10.2139/ssrn.4359318","title":"How to Cope with Malicious Federated Learning Clients:  an Unsupervised Learning-Based Approach","year":2023,"lang":"en","type":"article","venue":"SSRN Electronic Journal","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Ottawa","funders":"","keywords":"Unsupervised learning; Computer science; Federated learning; Artificial intelligence; Data science; Knowledge management; Machine learning; Psychology","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":[0.005621449,0.0009370035,0.00190187,0.001122544,0.002133186,0.003254288,0.004909765,0.004852945,0.00277093],"category_scores_gemma":[0.02011927,0.0006780241,0.001566887,0.001015233,0.002903738,0.00841015,0.005075973,0.003744974,0.001264874],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001407245,"about_ca_system_score_gemma":0.003457183,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003705728,"about_ca_topic_score_gemma":0.004298214,"domain_scores_codex":[0.9944755,0.001856442,0.0003021968,0.001147968,0.001473565,0.0007443409],"domain_scores_gemma":[0.9848534,0.005065127,0.001003787,0.005813483,0.002676306,0.0005879049],"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":[0.0009060873,0.0009856171,0.009902908,0.0002571067,0.0003640213,0.0005583812,0.001016363,0.4875868,0.01319559,0.06632452,0.01239389,0.4065087],"study_design_scores_gemma":[0.000009221261,0.00003708864,0.0002322873,0.00001268735,0.00001986282,0.0001324018,0.0001105754,0.969386,0.003056155,0.02612622,0.0008608458,0.00001659596],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01957265,0.00007968154,0.9749931,0.0009363007,0.00004929614,0.0001755344,0.00006603464,0.002081399,0.002046162],"genre_scores_gemma":[0.5940973,0.00009356876,0.3996634,0.000488839,0.00008913259,0.0001676386,0.0002179398,0.0003266452,0.004855584],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005621449,"threshold_uncertainty_score":0.02972943,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01895664919725563,"score_gpt":0.2503815268072572,"score_spread":0.2314248776100016,"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."}}