Zhiyuan Qi
About Me
I am a second-year M.S. student in the Department of Electronic Engineering at Tsinghua University, and I am majoring in Artificial Intelligence. I’m so fortunated to work with Dr.Yifan Shen and Dr.Jiateng Liu at UIUC, advised by Prof. Heng Ji.
My current research focuses on Spatial Intelligence for MLLM, specifically on How MLLMs better understand 3D space about positional, semantic and physical relationships and How to Improve MLLM’s Ability to Reason and Adopt Accurate Strategies integrated with 3D representations. I would be sincerely delighted to exchange ideas and seek collaboration opportunities with researchers and peers!
My long-term goal is to make the way MLLMs understand and represent 3D space and ultimately the real world both controllable and interpretable.
I’m currently seeking work opportunities! Please contact me for details!
Research Interests
- Multimodal Learning/MLLM: Spatial Intelligence, Unified Multimodal Understanding and Generation Models, LLM/VLM Agent, Agentic AI
- 3D Vision: World Model, 3DGS, Controllable Video Generation
News
- [Aug. 2026] Our paper RouteGraph-Mona is accepted by PRICAI 2026.
Publications
-
NAACL 2027
NAACL 2027, under review. -
PRICAI 2026
-
ICC 2025
IEEE International Conference on Communications (ICC), 2025. -
IEEE CM
IEEE Communication Magazine, 2025. -
AAAI 2027
AAAI Conference on Artificial Intelligence, 2027.
Patents
- “A Channel Environment Adaptive Joint Optimization Method and System for Sampling-Semantic-Channel Coding”, ZL 2024 1 1509807.8, Granted
- “A Joint Optimization Method, Device, and Medium for Sampling-Semantic in Point Clouds for Holographic Communications”, ZL 2025 1 1141478.0, Granted
- “A Method for Evaluating the Quality of a 3D Representation Experience for Immersive Communication”
Selected Awards
- Tsinghua University Comprehensive 1st Scholarship, 2025
- Outstanding Undergraduate Thesis Award, 2024
- National Scholarship (Awarded to the top 0.2% of students nationwide), 2022
- 20th China Graduate Electronic Design Competition, 2 nd Prize, 2025
- MCM (Mathematical Contest In Modeling)/ ICM (Interdisciplinary Contest In Modeling), Finalist Prize (Ranked in the top 1% of teams worldwide), 2021
Technical Skills
- Languages: Python, Matlab, C/C++
- Frameworks: Pytorch, Isaac-lab, MinkowskiEngine
- Tools: Unity 3D, Git, Conda, Cmake, Docker, Linux
Selected Services
Reviewer for CVPR 2026, NeurIPS 2026, PRICAI 2026, IEEE VTC Spring, IEEE ICME, IEEE COMMAG.
Powered by Jekyll and Minimal Light theme.