{"id":"W3185922622","doi":"10.1145/3462757.3466102","title":"Using transformers to improve answer retrieval for legal questions","year":2021,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"ca_institutions":"Thomson Reuters (Canada)","funders":"","keywords":"Computer science; Automatic summarization; Question answering; Transformer; Support vector machine; Artificial intelligence; Machine translation; Machine learning; Mean reciprocal rank; Natural language processing; Information retrieval; Engineering; Voltage","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":[0.002952623,0.001236966,0.0009123451,0.003582927,0.0006049302,0.001893156,0.001244694,0.00107066,0.0051309],"category_scores_gemma":[0.01200569,0.0002967437,0.000915022,0.001894516,0.0005131855,0.005900378,0.001617873,0.001428651,0.005135772],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001027087,"about_ca_system_score_gemma":0.001642243,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005917551,"about_ca_topic_score_gemma":0.009533649,"domain_scores_codex":[0.9984273,0.000464884,0.000135207,0.0004012663,0.0004144775,0.0001569354],"domain_scores_gemma":[0.996771,0.001579362,0.0001931053,0.0005265706,0.0007974452,0.0001324507],"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":[0.0007948104,0.0005750402,0.008975206,0.0004756986,0.000124805,0.0001876739,0.0005509446,0.02165562,0.03664748,0.01443779,0.0448891,0.8706859],"study_design_scores_gemma":[0.0001662188,0.0006743654,0.003652731,0.00004901555,0.0002110515,0.0004066449,0.0004024359,0.8884844,0.04952611,0.0248137,0.03155008,0.00006325021],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.244707,0.004974687,0.6518189,0.00237505,0.0006948955,0.0006757397,0.003388279,0.07396164,0.01740376],"genre_scores_gemma":[0.7643675,0.001212721,0.2111887,0.0007916379,0.000378874,0.0001995254,0.01135818,0.0008500522,0.009652843],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005917551,"threshold_uncertainty_score":0.01716453,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04129647095988698,"score_gpt":0.3094840678173827,"score_spread":0.2681875968574957,"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."}}