{"id":"W3167525829","doi":"10.18653/v1/2021.naacl-main.88","title":"DReCa: A General Task Augmentation Strategy for Few-Shot Natural Language Inference","year":2021,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":19,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Toyota Research Institute; Canadian Institute for Advanced Research","keywords":"Computer science; Inference; Task (project management); Natural language; Computational linguistics; Natural language processing; Artificial intelligence; Shot (pellet); Linguistics; Natural (archaeology); Engineering; Philosophy; History","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":[0.005063863,0.003107732,0.002873814,0.003871104,0.002407147,0.003123508,0.007569028,0.003236635,0.02028201],"category_scores_gemma":[0.0156576,0.002069689,0.003183037,0.002757453,0.000943043,0.00688936,0.005536667,0.006963027,0.01254831],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001125711,"about_ca_system_score_gemma":0.003409822,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01711721,"about_ca_topic_score_gemma":0.04610891,"domain_scores_codex":[0.9964993,0.001151463,0.0001733509,0.001305257,0.0005897228,0.0002809416],"domain_scores_gemma":[0.9932566,0.003271881,0.000122739,0.00203599,0.001032927,0.0002798232],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001205249,0.0008921895,0.001473179,0.0007654312,0.0007875129,0.0002943198,0.0005341782,0.02543506,0.01422188,0.01821315,0.1369156,0.7992622],"study_design_scores_gemma":[0.0001676491,0.0001309633,0.0006094222,0.000072362,0.0001888555,0.0002032528,0.0001286846,0.8989664,0.01073076,0.05768962,0.03100163,0.0001104955],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002769801,0.0007425443,0.950349,0.0002984469,0.0003709856,0.000281626,0.001941523,0.04094367,0.002302333],"genre_scores_gemma":[0.07830621,0.0005506086,0.8890172,0.0008746595,0.0005363679,0.001181851,0.0145204,0.005212181,0.009800526],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.02028201,"threshold_uncertainty_score":0.06785005,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0494776385202047,"score_gpt":0.3332134924141382,"score_spread":0.2837358538939335,"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."}}