{"id":"W4383604828","doi":"10.48550/arxiv.2307.02849","title":"NatLogAttack: A Framework for Attacking Natural Language Inference Models with Natural Logic","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Adversarial system; Computer science; Inference; Artificial intelligence; Spurious relationship; Natural language; Formalism (music); Natural language understanding; Non-monotonic logic; Natural (archaeology); Cognitive science; Machine learning; Psychology","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005053571,0.00106395,0.0007933767,0.001535275,0.001154913,0.003355376,0.003811158,0.002085277,0.007197229],"category_scores_gemma":[0.0127501,0.0008239151,0.002642458,0.0007462875,0.004823428,0.005242195,0.006143307,0.006071157,0.001276555],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002381323,"about_ca_system_score_gemma":0.002394746,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003052455,"about_ca_topic_score_gemma":0.004186824,"domain_scores_codex":[0.9963624,0.001595519,0.0001842474,0.0005665221,0.001046718,0.0002445528],"domain_scores_gemma":[0.9946621,0.003213107,0.0004012683,0.001227058,0.0002913368,0.0002052136],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001286386,0.00007692681,0.0004447792,0.0001686899,0.00009604486,0.000234856,0.0002849697,0.1589893,0.002734454,0.7908148,0.005121633,0.04090498],"study_design_scores_gemma":[0.00002473695,0.00004305885,0.00005203518,0.00003657335,0.00002752284,0.0001051753,0.0000307372,0.5105833,0.001738443,0.4768004,0.01053493,0.00002309536],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002346681,0.0001169404,0.9920658,0.0006482352,0.00004819124,0.0000987399,0.0001220229,0.001731188,0.002822227],"genre_scores_gemma":[0.2735765,0.0005586207,0.7155114,0.001167501,0.0001645732,0.0005750907,0.0005977126,0.0008148982,0.007033714],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007197229,"threshold_uncertainty_score":0.02672613,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1167031844234035,"score_gpt":0.2654593571134707,"score_spread":0.1487561726900672,"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."}}