I am a third-year PhD student at the School of Computing, National University of Singapore, where I am fortunate to be advised by Professor Angela Yao. I’m currently also a research intern at NVIDIA Efficient AI supervised by Dr. Yuyang Zhao and Professor Song Han. Previously, I am graduated in Turing Programme with Highest Distinction in Computer Science from the same faculty. My research is centered on developing efficient and controllable visual-centric generative models. I am particularly interested in understanding and modeling visual signals, as they are easy to obtain and can serve as observations of the real world with minimal inductive bias.

Before beginning my doctoral studies, I co-founded a local company Huijian Cloud in Guangzhou, where we developed AI agent solutions with real-world applicability, including PMS, IMS, and CRM platforms. The company has served clients across Mainland China, Malaysia, Thailand, and Australia. I was also a Fellow of MiraclePlus in 2021.

News

• 2026.01: Started my internship with the NVIDIA Efficient AI Team!
• 2026.01: Released InfoTok, a study on efficient video tokenization, in collaboration with NVIDIA Imaginary Lab.
• 2025.12: Released two papers on visual autoregressive models: VA-π and reAR.
• 2025.12: Released ReSlide, a tool for re-editing image-based PPTs generated by NotebookLM or Nano Banana.
• 2025.08: Honored to receive the Research Achievement Award at NUS!
• 2025.05: Honored to be NUS Search Committee Member!
• 2025.03: Released Conceptrol, a study on concept control for zero-shot personalized image generation.
• 2025.01: EvA accepted to ICLR 2025!
• 2024.09: AID accepted to NeurIPS 2024!
• 2024.05: Presented a research talk on generative models at the Huawei Singapore Harmony Software Technology Workshop.
• 2023.10: Honored to receive the Research Incentive Award at NUS!
• 2023.06: Honored to receive the NUS Outstanding Undergraduate Researcher Prize!
• 2023.02: PoseIG accepted to CVPR 2023!

Publications

(Preprint 2025) VA-π: Variational Policy Alignment for Pixel-Aware Autoregressive Generation
Xinyao Liao*, Qiyuan He*, Kai Xu, Xiaoye Qu, Yicong Li, Wei Wei, Angela Yao

(Preprint 2025) InfoTok: Adaptive Discrete Video Tokenizer via Information-Theoretic Compression
Haotian Ye*, Qiyuan He*, Jiaqi Han, Puheng Li, Jiaojiao Fan, Zekun Hao, Fitsum Reda, Yogesh Balaji, Huayu Chen, Sheng Liu, Angela Yao, James Zou, Stefano Ermon, Haoxiang Wang, Ming-Yu Liu

(Preprint 2025) reAR: Rethinking Visual Autoregressive Models via Token-wise Consistency Regularization
Qiyuan He, Yicong Li, Haotian Ye, Jinghao Wang, Xinyao Liao, Pheng-Ann Heng, Stefano Ermon, James Zou, Angela Yao

(Preprint 2025) Conceptrol: Concept Control of Zero-shot Personalized Image Generation
Qiyuan He, Angela Yao

(NeurIPS 2024) AID: Attention Interpolation of Text-to-Image Diffusion
Qiyuan He, Jinghao Wang, Ziwei Liu, Angela Yao

(ICLR 2025) EvA: Erasing Spurious Correlations with Activations
Qiyuan He, Kai Xu, Angela Yao

(CVPR 2023) Analyzing and Diagnosing Pose Estimation With Attributions
Qiyuan He*, Linlin Yang*, Kerui Gu, Qiuxia Lin, Angela Yao

Selected Awards

• NUS Search Commitee Member, 2025
• NUS Research Achievement Award, 2025
• NUS Research Incentive Award, 2023
• Outstanding Undergraduate Researcher Prize, 2023
• Certificate of Distinction in the AI Focus Area, 2022