{"id":"W3122996481","doi":"10.48550/arxiv.2101.09825","title":"Improving Few-Shot Learning with Auxiliary Self-Supervised Pretext Tasks","year":2021,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Computer science; Leverage (statistics); Pretext; Exploit; Machine learning; Complementarity (molecular biology); Artificial intelligence; Supervised learning; Task (project management); Training set; Semi-supervised learning; Class (philosophy); Labeled data","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003293866,0.0018796,0.002273682,0.001610097,0.001086053,0.001619611,0.003672132,0.002872861,0.003564817],"category_scores_gemma":[0.01485196,0.0007617656,0.001477871,0.001339176,0.001890527,0.005351128,0.003848647,0.004286597,0.002125189],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001315011,"about_ca_system_score_gemma":0.001422606,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004293228,"about_ca_topic_score_gemma":0.006632851,"domain_scores_codex":[0.9979799,0.0007805034,0.00008374749,0.000676517,0.0003207891,0.0001586221],"domain_scores_gemma":[0.9924259,0.004273957,0.0003553888,0.001841375,0.0007171865,0.0003862229],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0006169885,0.001285216,0.003168813,0.0006803787,0.0003301561,0.0002609871,0.0004872491,0.3969953,0.01672655,0.02191182,0.02394424,0.5335923],"study_design_scores_gemma":[0.00002355248,0.00006368788,0.0002333628,0.00001411004,0.00001312642,0.00003864663,0.00002453313,0.979745,0.003269135,0.0158098,0.0007509064,0.00001415414],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.03782988,0.0008639745,0.9533273,0.0004447508,0.0001344214,0.0001393779,0.0003546231,0.004569462,0.002336318],"genre_scores_gemma":[0.6262697,0.0004745668,0.3553075,0.001178558,0.0003971012,0.0003979451,0.004491589,0.001137422,0.01034561],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004293228,"threshold_uncertainty_score":0.01741987,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05132379531189979,"score_gpt":0.1775479464571536,"score_spread":0.1262241511452538,"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."}}