{"id":"W4385412128","doi":"10.1109/eurosp57164.2023.00023","title":"Reconstructing Individual Data Points in Federated Learning Hardened with Differential Privacy and Secure Aggregation","year":2023,"lang":"en","type":"article","venue":"","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto; Vector Institute","funders":"","keywords":"Differential privacy; Computer science; Data aggregator; Differential (mechanical device); Federated learning; Information privacy; Computer security; Data mining; Computer network; Artificial intelligence","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":[0.008490183,0.0007079954,0.001499483,0.0008513571,0.001303913,0.003056872,0.002132735,0.002046048,0.0009651919],"category_scores_gemma":[0.02112599,0.0006455086,0.001341979,0.001317558,0.003450369,0.006114656,0.007020304,0.003004711,0.0003020396],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001919573,"about_ca_system_score_gemma":0.001840271,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001263496,"about_ca_topic_score_gemma":0.0009240844,"domain_scores_codex":[0.9922706,0.003003706,0.000578366,0.001555138,0.001841505,0.0007506927],"domain_scores_gemma":[0.9859447,0.005224704,0.001117126,0.006662041,0.0007138379,0.0003376269],"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":[0.0009905001,0.0001957801,0.003703987,0.0001033057,0.0001745972,0.0005648365,0.000895597,0.680126,0.007995451,0.2101138,0.001698764,0.09343742],"study_design_scores_gemma":[0.00002868786,0.00005404448,0.0002128694,0.00001184763,0.00001454139,0.00009559913,0.00006126407,0.8926922,0.004505971,0.1016907,0.0006162768,0.00001595358],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05699427,0.00009335223,0.9409016,0.0004539545,0.00002561974,0.00005324348,0.00006677567,0.0006210002,0.0007902104],"genre_scores_gemma":[0.8906829,0.00006073268,0.1073343,0.0002095013,0.00002803871,0.00009624627,0.0001129986,0.00006262565,0.001412714],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008490183,"threshold_uncertainty_score":0.04490095,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05261085302713452,"score_gpt":0.2788809764056963,"score_spread":0.2262701233785618,"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."}}