Jiahui Li
School of Computing, University of Georgia
Athens, Georgia, USA · jl57095@uga.edu
Personal Website · Google Scholar · GitHub · LinkedIn

📄 Download this CV as PDF

🧬 Biography

Jiahui Li is a Ph.D. student in Computer Science at the University of Georgia, supervised by Dr. Fei Dou. His research focuses on evidence-grounded and explainable artificial intelligence for clinical diagnosis and health sensing, including multimodal clinical reasoning, physiological time-series modeling, and trustworthy evaluation. His work spans electrocardiography, photoplethysmography, ballistocardiography, bodyseismography, vision-language models, and instruction-tuned large language models.

📖 Education

• University of Georgia — Ph.D. in Computer Science, Aug. 2024 – Present. GPA 4.00/4.00.
• University College London — M.Sc. in Scientific and Data-Intensive Computing, 2021 – 2022. GPA 3.62/4.00.
• University of Liverpool — B.Sc. in Computer Science and Engineering, 2017 – 2021. GPA 3.92/4.00.

🔬 Research Experience

Graduate Research Assistant, School of Computing, University of Georgia — Aug. 2024 – Present
Advisor: Dr. Fei Dou

• DynamicDx (Lead): Built a reproducible benchmark of 71 video-grounded neurological consultations across 11 phenomenologies. Evaluation of five vision-language models localized the primary diagnostic bottleneck to evidence acquisition rather than downstream interpretation.
• EarlyDx (Lead): Designed an admission-anchored benchmark for open-ended, evidence-supported emergency-department diagnosis, testing whether generated diagnoses are grounded in retrievable clinical evidence.
• DeepArrhythmia (Lead): Developed segment-contextualized electrocardiogram classification with selective acquisition of waveform images, R-peak locations, and rhythm-morphology evidence, producing an auditable decision trace.
• BEACON-BP (Lead): Studied contactless blood-pressure estimation from bed-frame geophone sensing using invasive arterial-line measurements as the clinical reference.
• Peak-Detector / Peak-R1 (Lead): Built instruction-tuned large-language-model methods for explainable peak detection across ECG, PPG, BCG, and BSG without per-modality retraining.
• ADLGen (Co-Lead, co-first author): Synthesized symbolic, event-triggered smart-home sensor sequences for activity modeling and ambient assisted living.

Research Assistant, Hong Kong University of Science and Technology — 2023 – 2024
Supervisor: Prof. Dit-Yan Yeung

• Built MGTST, a multi-scale, cross-channel gated Transformer for multivariate long-term forecasting in high-dimensional time series.

📝 Publications

∗ equal contribution, † corresponding author. Author names in bold indicate myself.

First-Author Papers

• Jiahui Li, Ruili Fang, Zishuai Liu, Yutong Guo, Nan Yang, Wenzhan Song, Jin Lu, Fei Dou†. “EarlyDx: An Admission-Anchored Benchmark for Open-Ended Generation of Evidence-Supported ED-Encounter Diagnoses.” arXiv preprint, 2026.
• Jiahui Li, Yida Zhang, Zixuan Zeng, Jiayu Chen, Yingjian Song, Yin Xiao, Nishan Dong, Junjie Lu, Younghoon Kwon, Xiang Zhang, Jin Lu, Wenzhan Song, Fei Dou†. “Peak-Detector: Explainable Peak Detection via Instruction-Tuned Large Language Models in Physiological Signal.” ACM IMWUT (UbiComp) 2026.
• Weihang You∗, Hanqi Jiang∗, Jiahui Li∗ (co-first author), Zishuai Liu∗, Tianming Liu, Jin Lu, Fei Dou†. “ADLGen: Synthesizing Symbolic, Event-Triggered Sensor Sequences for Smart-Home Human Activity Modeling.” ACM SenSys 2026.
• Jiahui Li, Ruili Fang, Zishuai Liu, Wenzhan Song, Jin Lu, Fei Dou†. “DeepArrhythmia: Segment-Contextualized ECG Arrhythmia Classification via Selective Evidence Acquisition.” To appear at NeurIPS 2026.
• Jiahui Li, Yida Zhang, Zixuan Zeng, Jiayu Chen, Xiang Zhang, Jin Lu, Wenzhan Song, Fei Dou†. “Peak-R1: Instruction-Tuned Large Language Models for Robust J-Peak Detection.” Learning from Time Series for Health Workshop @ NeurIPS 2025.

Collaborative Papers

• Yuanyuan Zhang, Yida Zhang, Jiahui Li, Yuyan Wu, Fei Dou, Xiao Yin, Zhenlin An, Hae Young Noh, Wenzhan Song. “Physics-Constrained Deep Learning Model for Contactless Blood Pressure Monitoring from Triaxial Bodyseismography.” arXiv preprint, 2026.
• Haoran Lu, Luyang Fang, Ruidong Zhang, and others, including Jiahui Li. “Alignment and Safety in Large Language Models: Safety Mechanisms, Training Paradigms, and Emerging Challenges.” arXiv preprint, 2025.
• Zicheng Liu, Jiahui Li, Siyuan Li, Zelin Zang, Cheng Tan, Yufei Huang, Yajing Bai, Stan Z. Li. “GenBench: A Benchmarking Suite for Systematic Evaluation of Genomic Foundation Models.” arXiv preprint, 2024.

👨‍🏫 Teaching Assistant

• CSCI 1302: Software Development — Fall 2024 and Spring 2025, University of Georgia.
• CSCI 4470: Algorithms — Fall 2025 and Fall 2026, University of Georgia.