{"id":"W3129244551","doi":"10.48550/arxiv.2102.11072","title":"Obfuscation of Images via Differential Privacy: From Facial Images to General Images","year":2021,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Face recognition and analysis","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Carleton University","funders":"","keywords":"Obfuscation; Computer science; Differential privacy; Context (archaeology); Anonymity; Artificial intelligence; Pixel; Noise (video); Face (sociological concept); Generative grammar; Generative model; Machine learning; Image (mathematics); Computer vision; Computer security; Data mining","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.00009426614,0.0003438003,0.0005328234,0.0004049197,0.0001219677,0.0002673008,0.001510259,0.0002123132,0.0003207282],"category_scores_gemma":[0.00004481913,0.0003984329,0.0004581505,0.00065984,0.00009456734,0.0004635839,0.002332242,0.000334708,0.00007799851],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001146333,"about_ca_system_score_gemma":0.0001593071,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009828194,"about_ca_topic_score_gemma":0.00005189971,"domain_scores_codex":[0.9977968,0.0002115672,0.0003103491,0.001197728,0.0001832229,0.0003003382],"domain_scores_gemma":[0.9979774,0.00006615541,0.0003066419,0.001057179,0.0003872278,0.0002053503],"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":[0.0002464131,0.002757736,0.04227062,0.0007013115,0.003826158,0.001383717,0.00503963,0.1595133,0.6437,0.007747023,0.007732472,0.1250816],"study_design_scores_gemma":[0.002240256,0.0001418071,0.09384633,0.0004583311,0.001116928,0.00000623717,0.0004643453,0.4062637,0.47039,0.02186551,0.0004599905,0.002746616],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4401866,0.00003753143,0.5587376,0.0001857465,0.0003110003,0.0001174612,0.0001146144,0.00008566358,0.0002237124],"genre_scores_gemma":[0.9868177,0.0002116454,0.0113222,0.00007494991,0.0001382493,0.00000164484,0.0002291063,0.00001610716,0.001188438],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5474154,"threshold_uncertainty_score":0.9998468,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03945510140925789,"score_gpt":0.191435230540038,"score_spread":0.1519801291307801,"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."}}