{"id":"W4318719663","doi":"10.48550/arxiv.2301.11990","title":"Alignment with human representations supports robust few-shot learning","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Office of Naval Research; Natural Sciences and Engineering Research Council of Canada","keywords":"Adversarial system; Computer science; Artificial intelligence; Domain (mathematical analysis); Shot (pellet); Machine learning; Mathematics","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":[0.00465566,0.0009569078,0.001330662,0.0007996305,0.0007490327,0.002045609,0.001505256,0.001883665,0.001923878],"category_scores_gemma":[0.03168333,0.0005649927,0.0007211036,0.0007426225,0.002676015,0.005052324,0.003816812,0.002944467,0.0007250005],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009058094,"about_ca_system_score_gemma":0.001038281,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00213782,"about_ca_topic_score_gemma":0.001806169,"domain_scores_codex":[0.9960499,0.001818887,0.00012097,0.00103581,0.0006858365,0.0002885953],"domain_scores_gemma":[0.9834825,0.008532525,0.001842027,0.004846214,0.0007153731,0.0005812651],"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":[0.0009779836,0.0004785148,0.01382043,0.0003186672,0.000490582,0.0002903941,0.0007400888,0.7457253,0.02040804,0.07463571,0.004346355,0.1377679],"study_design_scores_gemma":[0.00002137572,0.0001800108,0.002219087,0.00001850322,0.00002824317,0.0001101099,0.0000969273,0.8888751,0.005440615,0.1020487,0.0009348949,0.00002639396],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2449325,0.0005677908,0.7454388,0.00145338,0.00008615128,0.00007403984,0.0002809883,0.001261695,0.005904674],"genre_scores_gemma":[0.97009,0.0001000211,0.02855152,0.0001621726,0.00004053419,0.00002518909,0.000291486,0.00006230622,0.000676616],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00465566,"threshold_uncertainty_score":0.02462178,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1226970041717997,"score_gpt":0.2397631785504393,"score_spread":0.1170661743786396,"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."}}