Tianqiu Zhang

I am currently a fourth-year Ph.D. student at Peking University, advised by Prof. Si Wu. My research focuses on world models and computational neuroscience.

My current research interests are:

• Latent World Models: brain-inspired world models, latent action models, structure learning.
• Brain-inspired Representation Learning: phase coding, multi-scale coding, equivariant map.
• Self-supervised Reinforcement Learning: goal-conditioned RL, skill learning, zero-shot generalization.
• Whole-brain Simulation: connectome-constrained models, sensorimotor closed-loop training.

Education

• Peking University, Ph.D. in Integrated Life Sciences (Physics), 2022–2027 (expected). Advised by Prof. Si Wu.
• Beijing University of Posts and Telecommunications, B.Eng. in Computer Science and Technology, 2018–2022.

My long-term academic goal is to learn a general embodied agent from vision and self-motion. I see world models as a language-independent way to learn general representations from continuous perceptual signals. I am also exploring whether bottom-up hierarchical abstraction from pure vision sequences can lead to compact, compositional, and transferable abstract structures.