{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004352312,0.0001132544,0.0001075734,0.000208657,0.0001851181,0.0005159792,0.000404984,0.00004488576,0.00012214],"category_scores_gemma":[0.0001201335,0.0001079702,0.0000578851,0.001204419,0.0000158052,0.000564249,0.0001889146,0.00007662655,0.0006376528],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006096113,"about_ca_system_score_gemma":0.0000622285,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004276757,"about_ca_topic_score_gemma":0.00001298933,"domain_scores_codex":[0.9986987,0.00008069223,0.0001374243,0.0003515464,0.0004211823,0.0003104964],"domain_scores_gemma":[0.9993919,0.0000323329,0.00004458803,0.0002717708,0.00007778726,0.0001816005],"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":[0.0001565093,0.0001433581,0.0006995297,0.00005588607,0.000112056,0.0004669542,0.04802091,0.04012175,0.0708909,0.2063651,0.2220991,0.410868],"study_design_scores_gemma":[0.002314252,0.0004016623,0.01637227,0.00002541405,0.00001867649,0.00003093834,0.009374255,0.7154845,0.01798603,0.003090408,0.2339568,0.0009447889],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06814743,0.00000965601,0.9136356,0.0112515,0.0006593573,0.0001970749,0.000002577379,0.0006974243,0.005399372],"genre_scores_gemma":[0.6767932,0.00002267747,0.1747099,0.004734844,0.000244733,0.00001883353,0.00003370314,0.0000271484,0.143415],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7389257,"threshold_uncertainty_score":0.819595,"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."}}