{"id":"W4399151655","doi":"10.48550/arxiv.2405.17878","title":"An Information Theoretic Evaluation Metric For Strong Unlearning","year":2024,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Advanced Statistical Methods and Models","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Instituto de Ciencias del Mar y Limnología, Universidad Nacional Autónoma de México; Institute for Catastrophic Loss Reduction","keywords":"Metric (unit); Computer science; Mathematics; Economics; Operations management","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001388974,0.0002331994,0.0002969765,0.0003672387,0.0001162327,0.00009092582,0.0002796431,0.0002322088,0.00007391227],"category_scores_gemma":[0.001414108,0.0002418341,0.0001473248,0.0003239555,0.00006144479,0.000307032,0.0002687298,0.0004965031,0.0000235909],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002886434,"about_ca_system_score_gemma":0.0001561561,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000009433735,"about_ca_topic_score_gemma":0.000004811607,"domain_scores_codex":[0.9986274,0.0002642264,0.0002717372,0.0004552021,0.0001367704,0.0002446542],"domain_scores_gemma":[0.9978893,0.0008881465,0.0002538578,0.0004767601,0.0003811119,0.0001108528],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00003623688,0.000023587,0.000006200117,0.0004244652,0.00004723879,0.000002557478,0.0002167552,0.1981612,0.000003986755,0.7880856,0.00004177876,0.01295029],"study_design_scores_gemma":[0.0001865945,0.00004797559,0.000006561432,0.00005931401,0.0003440531,3.432939e-7,0.0002009238,0.4865552,0.00001550179,0.5123812,0.00006426693,0.000138072],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03280946,0.00003673111,0.9622561,0.00001559943,0.0003405745,0.0009619611,0.0001226956,0.0001618263,0.003295027],"genre_scores_gemma":[0.8668548,0.00002327456,0.1326837,0.00001366408,0.00006476722,0.00001046263,0.000116976,0.00002815294,0.0002041621],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8340454,"threshold_uncertainty_score":0.9861708,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2659800695727216,"score_gpt":0.3548664914186651,"score_spread":0.08888642184594348,"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."}}