{"id":"W4405655508","doi":"10.32388/bojdxm","title":"Influence Functions for Scalable Data Attribution in Diffusion Models","year":2024,"lang":"en","type":"preprint","venue":"Qeios","topic":"Neural Networks and Applications","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute; University of Toronto","funders":"Engineering and Physical Sciences Research Council","keywords":"Attribution; Diffusion; Scalability; Computer science; Econometrics; Data science; Data mining; Psychology; Mathematics; Database; Physics; Thermodynamics; Social psychology","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":[0.0002323128,0.0001587369,0.0001691198,0.0001007052,0.0001134327,0.0002579922,0.001613128,0.0001558363,0.00000267665],"category_scores_gemma":[0.00001450013,0.0001488826,0.00005360745,0.0003778219,0.00001816853,0.0003216593,0.005075102,0.0004451838,0.00006329618],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007037902,"about_ca_system_score_gemma":0.0001203535,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001168424,"about_ca_topic_score_gemma":0.00008270871,"domain_scores_codex":[0.99843,0.00002112136,0.0002688585,0.0008796637,0.0001554448,0.0002449683],"domain_scores_gemma":[0.9979684,0.00007592063,0.00007422551,0.001749207,0.00007294877,0.00005923343],"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":[0.000006548209,0.000158077,0.0001085873,0.0002493405,0.00001890601,0.000004363942,0.0000950312,0.7633842,0.0006210453,0.1625281,0.03339456,0.03943128],"study_design_scores_gemma":[0.00007540744,0.000008686468,0.0002381853,0.0001438199,0.00001025611,0.000001677755,0.000001480091,0.8357263,0.00003606874,0.1529221,0.01069832,0.000137719],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05013705,0.0004739079,0.9435022,0.003645331,0.0005205038,0.0009105054,0.0003726679,0.0002249,0.0002129414],"genre_scores_gemma":[0.9824932,0.0001404777,0.01468653,0.0002138325,0.0002464593,0.0007255018,0.0007072023,0.0000188517,0.0007678984],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9323562,"threshold_uncertainty_score":0.6325751,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07695452809674291,"score_gpt":0.3101697495280635,"score_spread":0.2332152214313206,"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."}}