{"id":"W6910477211","doi":"10.48448/mccv-hq07","title":"Leveraging LLMs for Synthesizing Training Data Across Many Languages in Multilingual Dense Retrieval","year":2024,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Training set; Training (meteorology); Construct (python library); Language model; Data retrieval; Multilingualism; Labeled data; Document 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":[0.003971912,0.002175764,0.001313283,0.001753137,0.0007554603,0.001569702,0.002457273,0.001507306,0.005071099],"category_scores_gemma":[0.01436659,0.0008031161,0.001727987,0.001577424,0.0009559181,0.004511321,0.002911186,0.002336118,0.005060911],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001280442,"about_ca_system_score_gemma":0.00200884,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01205116,"about_ca_topic_score_gemma":0.02373083,"domain_scores_codex":[0.9977291,0.000897077,0.0001938016,0.000750814,0.0002628781,0.0001664117],"domain_scores_gemma":[0.9956093,0.002434689,0.0001773574,0.001098281,0.0005429331,0.0001374238],"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":[0.001266101,0.0009763017,0.007886644,0.002000343,0.000417652,0.0004619796,0.001061853,0.264351,0.03174885,0.0075842,0.05431717,0.6279278],"study_design_scores_gemma":[0.000296106,0.000555141,0.001327916,0.00009108366,0.000147371,0.0002554161,0.0003928762,0.9543149,0.01535436,0.01012779,0.01703655,0.0001005325],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2004745,0.006241535,0.6964484,0.00189698,0.0005126785,0.001257138,0.01114384,0.06932582,0.01269912],"genre_scores_gemma":[0.5502844,0.00089765,0.4014014,0.001612269,0.0002247938,0.001202307,0.03410174,0.002681846,0.0075935],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01205116,"threshold_uncertainty_score":0.02396202,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1674384594459624,"score_gpt":0.4350408651803633,"score_spread":0.2676024057344009,"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."}}