{"id":"W4389520055","doi":"10.18653/v1/2023.findings-emnlp.423","title":"How to Train Your Dragon: Diverse Augmentation Towards Generalizable Dense Retrieval","year":2023,"lang":"en","type":"article","venue":"","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":35,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Artificial intelligence; Relevance (law); Generative grammar; Encoder; Machine learning; Contrast (vision); Training set; Information retrieval","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":[0.002227218,0.001606158,0.001532471,0.0007655293,0.000483383,0.001079159,0.002312755,0.001689189,0.003748153],"category_scores_gemma":[0.006413148,0.0006553889,0.001023814,0.0006399296,0.001172636,0.003930887,0.002468384,0.002723644,0.002521056],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007373748,"about_ca_system_score_gemma":0.000932258,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005674206,"about_ca_topic_score_gemma":0.008945586,"domain_scores_codex":[0.9988493,0.0003541288,0.00005395596,0.0004223814,0.0001956441,0.0001245731],"domain_scores_gemma":[0.9980171,0.0009426564,0.00006847824,0.0006455678,0.0002303312,0.0000958192],"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":[0.001050758,0.0008115597,0.002708338,0.0007896031,0.0002802622,0.0005076181,0.0005885839,0.1908459,0.06479619,0.009991952,0.03785067,0.6897786],"study_design_scores_gemma":[0.0001179391,0.0004024965,0.000524087,0.00003312259,0.00005832281,0.0003423656,0.000110457,0.962905,0.02079429,0.00870973,0.005956622,0.00004560097],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1148751,0.002472156,0.850746,0.0008746054,0.0002512236,0.0004142957,0.001110903,0.0233778,0.005877964],"genre_scores_gemma":[0.5663275,0.0006207972,0.4141015,0.002049884,0.0002022241,0.000371854,0.00517373,0.001270462,0.009882098],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005674206,"threshold_uncertainty_score":0.01253879,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07580167410820143,"score_gpt":0.301941547100988,"score_spread":0.2261398729927866,"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."}}