{"id":{"repo_id":"uoit","oai_identifier":"oai:ontariotechu.scholaris.ca:10155/2098"},"canonical_url":"https://search.dev.ndltd.org/etd/uoit/oai:ontariotechu.scholaris.ca:10155/2098","repository":{"repo_id":"uoit","name":"Ontario Institute of Technology","base_url":"https://ontariotechu.scholaris.ca/server/oai/request"},"display":{"title":"ClinicalTrACE: a self-correcting agent with interpretable uncertainty for clinical question answering","abstract":"Answering precise questions about patient records requires retrieving events that satisfy type, temporal, and content constraints simultaneously, a multi-constraint satisfaction problem that embedding-based systems cannot solve. We introduce ClinicalTrACE, a self-correcting agent that retrieves through explicit structured queries with no fine-tuning and no task-specific training data, achieving 94.1% accuracy versus 76.5% for RAG and 55.3% for a fine-tuned baseline trained on 400K examples. Ablation shows that retrieval design drives this gain: categorical constraints alone contribute +12.6%, while scaling from 3B to 14B parameters adds only +5.1%. We also develop TrACE+, an uncertainty framework that predicts errors from ClinicalTrACE’s observable execution trace. A domain-agnostic variant achieves 0.83 AUROC with stable calibration (ECE = 0.060) and transfers across hospital systems; a domain-aware variant reaches 0.86 AUROC and 97.9% accuracy at 75% coverage at the cost of portability, revealing a clear generalization-discrimination tradeoff.","abstract_html":"Answering precise questions about patient records requires retrieving events that satisfy type, temporal, and content constraints simultaneously, a multi-constraint satisfaction problem that embedding-based systems cannot solve. We introduce ClinicalTrACE, a self-correcting agent that retrieves through explicit structured queries with no fine-tuning and no task-specific training data, achieving 94.1% accuracy versus 76.5% for RAG and 55.3% for a fine-tuned baseline trained on 400K examples. Ablation shows that retrieval design drives this gain: categorical constraints alone contribute +12.6%, while scaling from 3B to 14B parameters adds only +5.1%. We also develop TrACE+, an uncertainty framework that predicts errors from ClinicalTrACE’s observable execution trace. A domain-agnostic variant achieves 0.83 AUROC with stable calibration (ECE = 0.060) and transfers across hospital systems; a domain-aware variant reaches 0.86 AUROC and 97.9% accuracy at 75% coverage at the cost of portability, revealing a clear generalization-discrimination tradeoff.","abstract_has_math":false,"creators":["Wadie, Peter"],"institution":"University of Ontario Institute of Technology","degree_name":"Master of Applied Science (MASc)","degree_level":null,"degree_discipline":"Software Engineering","degree_department":null,"school":null,"contributors":[],"advisors":["Elgazzar, Khalid","Alwidian, Sanaa"],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026-04-01","date_published":"2026-04-01","updated_at":"2026-07-24T05:35:22Z","subjects":[],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/10155/2098","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Elgazzar, Khalid","Alwidian, Sanaa"]},{"key":"dc:creator","label":"Author","values":["Wadie, Peter"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2026-04-28T20:12:39Z"]},{"key":"dc:date.issued","label":"Date","values":["2026-04-01"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Software Engineering"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Master of Applied Science (MASc)"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Ontario Institute of Technology"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/10155/2098"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Answering precise questions about patient records requires retrieving events that satisfy type, temporal, and content constraints simultaneously, a multi-constraint satisfaction problem that embedding-based systems cannot solve. We introduce ClinicalTrACE, a self-correcting agent that retrieves through explicit structured queries with no fine-tuning and no task-specific training data, achieving 94.1% accuracy versus 76.5% for RAG and 55.3% for a fine-tuned baseline trained on 400K examples. Ablation shows that retrieval design drives this gain: categorical constraints alone contribute +12.6%, while scaling from 3B to 14B parameters adds only +5.1%. We also develop TrACE+, an uncertainty framework that predicts errors from ClinicalTrACE’s observable execution trace. A domain-agnostic variant achieves 0.83 AUROC with stable calibration (ECE = 0.060) and transfers across hospital systems; a domain-aware variant reaches 0.86 AUROC and 97.9% accuracy at 75% coverage at the cost of portability, revealing a clear generalization-discrimination tradeoff."]},{"key":"dc:title","label":"Title","values":["ClinicalTrACE: a self-correcting agent with interpretable uncertainty for clinical question answering"]}]}],"canonical_facts":{"dc:contributor.advisor":["Elgazzar, Khalid","Alwidian, Sanaa"],"dc:creator":["Wadie, Peter"],"dc:date.accessioned":["2026-04-28T20:12:39Z"],"dc:date.issued":["2026-04-01"],"dc:description.abstract":["Answering precise questions about patient records requires retrieving events that satisfy type, temporal, and content constraints simultaneously, a multi-constraint satisfaction problem that embedding-based systems cannot solve. We introduce ClinicalTrACE, a self-correcting agent that retrieves through explicit structured queries with no fine-tuning and no task-specific training data, achieving 94.1% accuracy versus 76.5% for RAG and 55.3% for a fine-tuned baseline trained on 400K examples. Ablation shows that retrieval design drives this gain: categorical constraints alone contribute +12.6%, while scaling from 3B to 14B parameters adds only +5.1%. We also develop TrACE+, an uncertainty framework that predicts errors from ClinicalTrACE’s observable execution trace. A domain-agnostic variant achieves 0.83 AUROC with stable calibration (ECE = 0.060) and transfers across hospital systems; a domain-aware variant reaches 0.86 AUROC and 97.9% accuracy at 75% coverage at the cost of portability, revealing a clear generalization-discrimination tradeoff."],"dc:identifier.uri":["https://hdl.handle.net/10155/2098"],"dc:language.iso":["en"],"dc:title":["ClinicalTrACE: a self-correcting agent with interpretable uncertainty for clinical question answering"],"dc:type":["Thesis"],"thesis:degree_discipline":["Software Engineering"],"thesis:degree_name":["Master of Applied Science (MASc)"],"thesis:institution_name":["University of Ontario Institute of Technology"]},"updated_at":"2026-07-24T05:35:22Z"}