{"id":"W3217374154","doi":"10.1109/jiot.2021.3131258","title":"A Fine-Grained Differentially Private Federated Learning Against Leakage From Gradients","year":2021,"lang":"en","type":"article","venue":"IEEE Internet of Things Journal","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":42,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of New Brunswick","funders":"Basic and Applied Basic Research Foundation of Guangdong Province; China Postdoctoral Science Foundation; National Natural Science Foundation of China","keywords":"Computer science; Differential privacy; Adversary; Leakage (economics); Federated learning; Deep learning; Information privacy; Private information retrieval; Data mining; Artificial intelligence; Distributed computing; Computer security","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":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","open_science"],"consensus_categories":["open_science"],"category_scores_codex":[0.0005025031,0.0002754217,0.0004270594,0.0002229598,0.000180976,0.0009116438,0.014927,0.0001759718,0.00007339358],"category_scores_gemma":[0.01043752,0.0002551141,0.0002039394,0.0003896625,0.0001007944,0.001234676,0.02035228,0.001399762,0.00003073459],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001256335,"about_ca_system_score_gemma":0.0001122752,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007526142,"about_ca_topic_score_gemma":0.00002079604,"domain_scores_codex":[0.9973632,0.0002457421,0.0007275831,0.0005474085,0.000631492,0.000484575],"domain_scores_gemma":[0.9965907,0.0002019139,0.0006885567,0.002064285,0.0003111202,0.0001433944],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0001477727,0.0006777138,0.01321323,0.00008981408,0.001267058,0.003459772,0.003531019,0.000210695,0.6719183,0.001230692,0.1532791,0.1509749],"study_design_scores_gemma":[0.002081129,0.0002844911,0.003130302,0.001162536,0.00004344315,0.0004666187,0.00008672389,0.3185252,0.5697817,0.1001006,0.00363082,0.0007064435],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6015998,0.000138685,0.3931762,0.003170124,0.001309326,0.00004504386,0.000004435591,0.0002597504,0.0002966071],"genre_scores_gemma":[0.88041,0.00007413349,0.1188085,0.0002451319,0.00009620211,0.000002241522,0.00001915019,0.00002442292,0.0003202015],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3183146,"threshold_uncertainty_score":0.9999901,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02128934842707645,"score_gpt":0.2508348126022749,"score_spread":0.2295454641751984,"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."}}