{"id":"W4412691112","doi":"10.22260/isarc2025/0088","title":"RAG-Enhanced Safety Information Retrieval for Construction: Integration of Large Language Models with Domain-Specific Information","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ... ISARC","topic":"Occupational Health and Safety Research","field":"Health Professions","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"University of Alberta","keywords":"Computer science; Domain (mathematical analysis); Information retrieval; Information integration; Natural language processing; Information model; Artificial intelligence; Data mining; Software engineering; Mathematics","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002317322,0.00100704,0.0008233658,0.001539863,0.0003677257,0.001564863,0.001315566,0.001086944,0.005934286],"category_scores_gemma":[0.006698046,0.0004104381,0.001031593,0.000814123,0.000526168,0.002912691,0.001715969,0.0009530977,0.005696876],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000627662,"about_ca_system_score_gemma":0.0009035048,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003293795,"about_ca_topic_score_gemma":0.003098819,"domain_scores_codex":[0.998678,0.0005595459,0.0001132824,0.0003124581,0.0002563463,0.00008023882],"domain_scores_gemma":[0.996886,0.001620053,0.0002568006,0.0006137294,0.0005313908,0.00009201976],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00109436,0.0006659512,0.004490812,0.0009301122,0.0002013641,0.0006324428,0.001584176,0.0424207,0.1173844,0.006259644,0.02108369,0.8032524],"study_design_scores_gemma":[0.0001222893,0.0005785467,0.002984746,0.00007480101,0.0001505914,0.0006427398,0.0005467264,0.8888232,0.06900875,0.008098361,0.02881835,0.0001508735],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.06135467,0.0005647787,0.8636692,0.0005357378,0.00009075928,0.0004999989,0.001659967,0.06830218,0.00332271],"genre_scores_gemma":[0.3963106,0.0003182821,0.591746,0.0004352291,0.00007492892,0.0003501868,0.005015283,0.001062907,0.004686652],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005934286,"threshold_uncertainty_score":0.01985222,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02371043577322171,"score_gpt":0.368119619173015,"score_spread":0.3444091833997933,"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."}}