Hello! I am Yuheng Zhao (่ตตๅฎๆ), a third-year PhD candidate in the School of Data Science at Fudan University, supervised by Prof. Siming Chen in FDUVIS Lab. I am also a visiting PhD student at the University of Edinburgh, supervised by Prof. Benjamin Bach in VisHub(https://vishub.net/). I received my Masterโs degree with First Class Honors from The Hong Kong University of Science and Technology (HKUST) in 2021, majoring in Data-driven Modeling under the supervision of Prof. Wong Kwok Yee Michael. I also completed a research internship at Huawei, advised by Dr. Yu Zhang.
My research interest lies in harnessing the confluence of Artificial Intelligence, Human-Computer Interaction and Visualization techniques, to facilitate data expression and comprehension. Lately, my primary focus has been on mixed-initiative data analysis, LLM reasoning and task planning, UI agent.
๐ I am actively looking for a Postdoc position starting in 2026! Please feel free to contact me if you have any opportunities! ๐
๐ Projects
We propose a proactive visual analytics frameowork that utilizes LLM-powered UI agent to help users with context-aware assistance proactively. The agent can conduct interactive and visual reasoning in visualization system.
LightVA is a lightweight VA framework through human-agent collaboration based on LLM agent-based task planning and execution.
LEVA is a framework that integrates LLMs to enhance userโs visual analytics workflow in three pivotal stages: onboarding, exploration, and summarization.
๐ Publications
The ACM Symposium on User Interface Software and Technology (UIST), 2025. ProactiveVA: Proactive Visual Analytics with LLM-Based UI Agent
IEEE Transactions on Visualization and Computer Graphics (VIS), 2025. SceneLoom: Communicating Data with Scene Context
IEEE Transactions on Visualization and Computer Graphics (VIS), 2025. SmartMLVs: LLM Enabled Multiple Linked Views Generation for Interactive Visualization
IEEE Pacific Visualization Conference (PacificVis), 2025.
Best Paper Honorable Mention Social Media Island: Interactive User Profiling and Information Diffusion Exploration with 3D Visual Metaphors
IEEE Pacific Visualization Conference (PacificVis), 2025. LightVA: Lightweight Visual Analytics with LLM Agent-Based Task Planning and Execution
IEEE Transactions on Visualization and Computer Graphics (TVCG), 2024. PCLVis: Visual Analytics of Process Communication Latency in Large-Scale Simulation
Computational Visual Media (CVMJ), 2024. LEVA: Using Large Language Models to Enhance Visual Analytics
IEEE Transactions on Visualization and Computer Graphics (TVCG), 2024. OldVisOnline: Curating a Dataset of Historical Visualizations
IEEE Transactions on Visualization and Computer Graphics (VIS), 2023. Visual Exploration of the Reading Comprehension Question Answering Model for Improving the Interpretability
IEEE Transactions on Visualization and Computer Graphics (TVCG), 2023. ContextWing: Pair-wise Visual Comparison for Evolving Sequential Patterns of Contexts in Social Media Data Streams
Proceedings of ACM SIGCHI Conference on Computer-Supported Cooperative Work and Social Computing (CSCW), 2023. Metaverse: Perspectives from graphics, interactions and visualization
Visual Informatics, 2022. Mixed-Initiative Visual Exploration of Social Media Text and Events
Proceedings of IEEE VIS, IEEE VAST Challenge 2021 - Award for Strong Human-in-the-Loop Analysis Methodology, 2021.
Award for Strong Human-in-the-Loop Analysis Methodology A Visualization System for Spatio-temporal Situation Awareness
Proceedings of IEEE VIS, IEEE VAST Challenge 2021 - Award for Outstanding Comprehensive MC2 Submission., 2021.
Outstanding Comprehensive Solution Award