Pengze Li

Pengze Li

Auto-generated profile, based on publicly available information

Pengze Li

AI & Informatics, Mayo Clinic, 4500 San Pablo Road, Jacksonville, 32224, FL, USA

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications

22 April 2025
Verified
Conference Paper
Open access
MSMingxuan Song
Mingxuan Song
PLPengze Li
Pengze Li
BZBohan Zhou
Bohan Zhou
SYShenglin Yin
Shenglin Yin
ZXZhen Xiao
Zhen Xiao
JLJieyi Long
Jieyi Long

24 October 2025 in Journal of Biomedical Informatics
Verified
Journal Article
Closed access
JLJianfu Li
Jianfu Li
YLYiming Li
Yiming Li
ZSZenan Sun
Zenan Sun
EYEvan Yu
Evan Yu
AAAhmed M. Abdelhameed
Ahmed M. Abdelhameed
CTCui Tao
Cui Tao

6 January 2026 in Computational and Structural Biotechnology Journal
Verified
Journal Article
Open access
PLPengze Li
Pengze Li
YHYutong Hu
Yutong Hu
JLJianfu Li
Jianfu Li
GGGarit Gemeinhardt
Garit Gemeinhardt
FLFang Li
Fang Li
CTCui Tao
Cui Tao
Digital twins have emerged as a paradigm in precision and personalized medicine, enabling data-driven modeling of individuals to support tailored interventions. While most existing work focuses on patient-oriented twins, little attention has been given to modeling the provider’s role, particularly in clinical communication. In this study, we present GRACE (Generalized RAG-Enhanced Conversation Framework), a framework for constructing a provider digital twin (ProDT) that emulates key aspects of clinicians’ communicative and cognitive behavior. GRACE integrates three modules: a physician-informed dialog script generation and optimization module for provider-patterned conversation, a Retrieval-Augmented Generation (RAG) pipeline for factual grounding and timely knowledge updating, and an LLM-based conversational interface that enables interactive, context-aware exchanges. Using HPV vaccination counseling as a representative use case, GRACE was evaluated with HealthBench and a structured user study involving clinician feedback. The results demonstrate its feasibility, trustworthiness, and adaptability for proactive provider–patient communication, marking a conceptual step toward safe, scalable, and cognitively grounded digital twins in healthcare.

30 March 2025
Verified
Conference Paper
Open access
LPLichen Pan
Lichen Pan
JLJuncheng Liu
Juncheng Liu
YFYongquan Fu
Yongquan Fu
JYJinhui Yuan
Jinhui Yuan
RZRongkai Zhang
Rongkai Zhang
ZXZhen Xiao
Zhen Xiao

10 November 2025 in CIKM: Conference on Information and Knowledge Management
Verified
Conference Paper
Open access
DJDong Hyun Jeon
Dong Hyun Jeon
LZLijing Zhu
Lijing Zhu
HLHaifang Li
Haifang Li
PLPengze Li
Pengze Li
JFJingna Feng
Jingna Feng
SNShuteng Niu
Shuteng Niu
Temporal Graph Neural Networks (TGNNs) have become indispensable for analyzing dynamic graphs in critical applications such as social networks, communication systems, and financial networks. However, the robustness of TGNNs against adversarial attacks, particularly sophisticated attacks that exploit the temporal dimension, remains a significant challenge. Existing attack methods for Spatio-Temporal Dynamic Graphs (STDGs) often rely on simplistic, easily detectable perturbations (e.g., random edge additions/deletions) and fail to strategically target the most influential nodes and edges for maximum impact. We introduce the High Impact Attack (HIA), a novel restricted black-box attack framework specifically designed to overcome these limitations and expose critical vulnerabilities in TGNNs. HIA leverages a data-driven surrogate model to identify structurally important nodes (central to network connectivity) and dynamically important nodes (critical for the graph’s temporal evolution). It then employs a hybrid perturbation strategy, combining strategic edge injection (to create misleading connections) and targeted edge deletion (to disrupt essential pathways), maximizing TGNN performance degradation. Importantly, HIA minimizes the number of perturbations to enhance stealth, making it more challenging to detect. Comprehensive experiments on five real-world datasets and four representative TGNN architectures (TGN, JODIE, DySAT, and TGAT) demonstrate that HIA significantly reduces TGNN accuracy on the link prediction task, achieving up to a 35.55% decrease in Mean Reciprocal Rank (MRR) – a substantial improvement over state-of-the-art baselines. These results highlight fundamental vulnerabilities in current STDG models and underscore the urgent need for robust defenses that account for both structural and temporal dynamics. Code and Data are available at https://github.com/ryandhjeon/hia.