Bowen Zheng

I am a second-year PhD student at The Chinese University of Hong Kong, Shenzhen, and a research intern at Tencent Hunyuan. My research focuses on video generation and autoregressive visual modeling: how models use temporal context, how visual representations shape generation, and how to make generation more efficient. I am particularly interested in extending these capabilities toward long-horizon, interactive, and real-time world simulation. My earlier work explored one-step diffusion generation, in-context learning, and efficient inference in language models.