Article Per Year (5 Year)
p-Index From 2021 - 2026
0.408
P-Index
This Author published in this journals
All Journal Journal of Technology Informatics and Engineering International Journal of Graphic Design
Zijian Cao
Electrical and Computer Engineering, University of Wisconsin-Milwaukee, WI, USA
Arts Computer Science & IT

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search

 1 
Offline Routing for Health Language Systems with Clarification and Escalation Zijian Cao; Justin Gao
Journal of Technology Informatics and Engineering Vol. 4 No. 2 (2025): AUGUST | JTIE : Journal of Technology Informatics and Engineering
Publisher : University of Science and Computer Technology

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.51903/jtie.v4i2.591

Abstract

Health-language systems require a control decision before response generation: answer with the available information, request clarification, or escalate for professional review. This study compared eight offline routing policies on HealthBench using researcher-defined actions mapped from physician-agreed categories. The implementations used TF-IDF logistic regression, bootstrap ensembles, stacked classification, and deterministic decision rules. A cohort of 2,601 dialogues provided disjoint development partitions, a 323-dialogue in-distribution test, and a locked 453-dialogue Hard test. Six policies used dialogue-derived information; two also accessed case-specific physician rubrics. Under this privileged-information condition, the rubric-augmented router achieved the highest accuracy, 69.97% in-distribution and 65.34% on Hard. The privileged cascade had the lowest severity-weighted routing loss, 0.1564 and 0.2851. Removing its rubric stage increased Hard loss to 0.3324 and unsafe direct-answer omission from 36.72% to 43.28%. Nevertheless, the privileged cascade directly escalated only 9/16 urgent cases; 6/16 were routed to clarification and 1/16 to an answer. Rubric access improves offline discrimination, but the strongest results do not establish an equivalent deployable verifier. The framework supports controlled prototyping of pre-generation routing, with substantial residual errors and no evaluation of generated medical responses or clinical outcomes.
Visualizing Uncertainty in AI-Assisted Canine Acute-Pain Assessment: Dog-Grouped Risk Modeling and Information-Equivalent Interface Prototypes Chole Zhao; Evelyn Liu; Zijian Cao
International Journal of Graphic Design Vol. 3 No. 1 (2025): May | IJGD: International Journal of Graphic Design
Publisher : University of Science and Computer Technology

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.51903/ijgd.v3i1.4272

Abstract

Canine acute-pain assessment combines behavioral observation, clinical context, and professional judgment, yet a model probability can be mistaken for a clinical pain score when the two lack a clear visual hierarchy. This study linked retrospective prediction-model findings to three information-equivalent interface prototypes. The structured database contained 594 observations from 184 dogs; 114 complete acute-pain observations from 65 dogs formed the primary cohort, with the short-form Glasgow Composite Measure Pain Scale (CMPS-SF) ≥ 6/24 as the reference outcome. A hypothetical pre-score feature set excluded direct score components, although the database did not establish recording order. Four calibrated models were evaluated with five repetitions of four-fold dog-grouped nested cross-validation. Logistic regression achieved an area under the receiver operating characteristic curve (AUROC) of .627 (95% CI [.474, .755]), an area under the precision–recall curve (AUPRC) of .476 [.317, .660], and Brier score .210 [.174, .249]. Sensitivity at 0.50 was .189, missing 30 of 37 threshold-positive observations. Full-development bootstrap fits supplied 90% model-stability ranges, distinct from patient-outcome intervals. Exploratory routing reduced Brier score but also reduced AUROC and changed retained event prevalence. An auxiliary vocabulary analysis identified 169 descriptors in 52 grouped rows. A separate transit-display analysis of 12,331 records from 299 participant identifiers did not establish an average payoff or duration advantage over a no-uncertainty display. The numeric dashboard, visual scale, and progressive evidence card retained identical information while varying hierarchy and disclosure. Their structural audit establishes content equivalence, not clinical effectiveness. The design contribution connects model limitations to visible range definitions, separate clinical-reference labels, missing-context cues, and a persistent structured-assessment action.
Co-Authors Chole Zhao Evelyn Liu Justin Gao