{"id":"W4396611084","doi":"10.21203/rs.3.rs-4320842/v1","title":"Enhancing Practicality and Efficiency of Deepfake Detection.","year":2024,"lang":"en","type":"preprint","venue":"Research Square","topic":"Digital Media Forensic Detection","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"Computer Research Institute of Montréal","funders":"","keywords":"Computer science; Key (lock); Inference; Convolutional neural network; Artificial intelligence; Enhanced Data Rates for GSM Evolution; Image (mathematics); Machine learning; Core (optical fiber); 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":[],"consensus_categories":[],"category_scores_codex":[0.002919128,0.0001823239,0.0002747963,0.0006217667,0.0001105428,0.0005649235,0.000617174,0.0002482641,0.000006001198],"category_scores_gemma":[0.002802265,0.0001675035,0.0001024339,0.0009771567,0.000295536,0.0002610477,0.004815262,0.001727152,0.00007542886],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002026797,"about_ca_system_score_gemma":0.0004404928,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001771948,"about_ca_topic_score_gemma":0.0001363126,"domain_scores_codex":[0.996545,0.000330654,0.000387905,0.0008621294,0.001404159,0.0004701834],"domain_scores_gemma":[0.9972607,0.0008755008,0.0001051053,0.0009396631,0.0006238944,0.0001951672],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00006151096,0.0003843472,0.0001420729,0.01080955,0.0001381639,0.000189322,0.004416927,0.0001201572,0.02018197,0.04352297,0.0002667264,0.9197663],"study_design_scores_gemma":[0.0003541911,0.00130838,0.002683342,0.003516125,0.00003720065,0.0001472518,0.0007238609,0.08578201,0.495773,0.4061896,0.002710586,0.0007744465],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6089955,0.00273574,0.3674522,0.001321438,0.003525923,0.00174889,0.00003193511,0.0006058487,0.01358247],"genre_scores_gemma":[0.9961371,0.00006688945,0.003411315,0.00000603616,0.0001354746,0.00007938588,0.000002015133,0.00001838497,0.0001434245],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9189919,"threshold_uncertainty_score":0.7503712,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04351096778690593,"score_gpt":0.3783818313598122,"score_spread":0.3348708635729063,"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."}}