{"id":"W3217532475","doi":"10.21203/rs.3.rs-1005694/v1","title":"Federated Learning and Differential Privacy for Medical Image Analysis","year":2021,"lang":"en","type":"preprint","venue":"Research Square","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":104,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Guelph; University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada; Government of Ontario; Compute Canada","keywords":"Computer science; Confidentiality; Differential privacy; Federated learning; Machine learning; Artificial intelligence; Support vector machine; Scale (ratio); Health care; Data science; Data mining; Computer security","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009630846,0.000498727,0.0009965118,0.000922355,0.0008284387,0.002541044,0.001947516,0.00151535,0.001462139],"category_scores_gemma":[0.02968016,0.00035778,0.0008702165,0.001588518,0.002564546,0.004231123,0.003138137,0.002173683,0.0004387385],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001827059,"about_ca_system_score_gemma":0.001569225,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008078137,"about_ca_topic_score_gemma":0.0006397561,"domain_scores_codex":[0.9941649,0.003004966,0.0002733545,0.0008889379,0.001323861,0.0003439671],"domain_scores_gemma":[0.9788245,0.01095331,0.001311398,0.00737821,0.001196163,0.0003364356],"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":[0.001177492,0.0004823973,0.007991262,0.00018796,0.0001906565,0.000445775,0.0004296512,0.5464118,0.01019538,0.1566964,0.005749688,0.2700414],"study_design_scores_gemma":[0.00002534578,0.00007115379,0.0005461373,0.00001582975,0.00001503878,0.0001740154,0.00004204123,0.8599985,0.007819648,0.1297802,0.001500441,0.00001159445],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.03022569,0.0003328939,0.9662781,0.001117573,0.00004963831,0.00005025141,0.0001575379,0.0006221283,0.001166118],"genre_scores_gemma":[0.8311508,0.0002471966,0.1661579,0.0004102949,0.0001001682,0.0001032401,0.0002677999,0.00007170257,0.001490797],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.009630846,"threshold_uncertainty_score":0.05093336,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0609896629642907,"score_gpt":0.407527223854721,"score_spread":0.3465375608904303,"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."}}