{"id":"W4389518777","doi":"10.18653/v1/2023.emnlp-main.257","title":"Transductive Learning for Textual Few-Shot Classification in API-based Embedding Models","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"École de Technologie Supérieure","funders":"Grand Équipement National De Calcul Intensif","keywords":"Computer science; Embedding; Artificial intelligence; Natural language processing; Shot (pellet); Machine learning; Chemistry","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":[0.003818822,0.001429125,0.002535513,0.001866513,0.0009190416,0.002385745,0.004016658,0.002607649,0.003873628],"category_scores_gemma":[0.01314055,0.0008309227,0.001350738,0.001739497,0.001224642,0.006065005,0.002725222,0.004090856,0.002411962],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001720772,"about_ca_system_score_gemma":0.000800672,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004337281,"about_ca_topic_score_gemma":0.006644165,"domain_scores_codex":[0.9977263,0.001069644,0.0001343169,0.0006155347,0.0002467835,0.00020746],"domain_scores_gemma":[0.9885327,0.008971511,0.0004006704,0.0009772776,0.0008075569,0.0003103674],"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":[0.0008826593,0.001167562,0.004422606,0.0006290863,0.0003597216,0.0002202179,0.0006510703,0.373577,0.00265372,0.03263583,0.02646633,0.5563342],"study_design_scores_gemma":[0.00001046838,0.00003481363,0.0001714656,0.00001543532,0.00001294707,0.00001650968,0.00002748847,0.9770098,0.0002625978,0.02199563,0.0004329417,0.000009927999],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06309764,0.003862728,0.9241649,0.001501686,0.0003010563,0.0002264252,0.001014729,0.002816889,0.003013798],"genre_scores_gemma":[0.8531986,0.001034324,0.127023,0.0008274394,0.0007976911,0.0006740938,0.004715159,0.000667127,0.01106251],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004337281,"threshold_uncertainty_score":0.02019614,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1290711266498333,"score_gpt":0.3314435657339576,"score_spread":0.2023724390841243,"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."}}