'Diffusion Models' Trends

Fuel your creativity, frame your story. Create, edit, and remix images and videos. AI video and image generator for creators. Text to video, AI headshots, consistent characters. Seedance 2.5 coming soon — an official ByteDance partnership. Powered by Nano Banana 2, Veo 3.1, Kling 3.0, Seedance 2.0 & Seedance 2.5.

Diffusion models for reinforcement learning and planning

Heng Dong is a researcher at ByteDance, working on embodied intelligence and multi-modal foundation models for robotics. He aims to build superhuman AI agents by combining learning from interaction—such as reinforcement learning, robot control and design, and multi-agent systems—with modern AI models, including large language models (LLMs), vision-language models (VLMs), diffusion models, and flow-based models. He received his Ph.D. in Computer Science from the Institute for Interdisciplinary Information Sciences (IIIS), Tsinghua University, under the supervision of Prof. Chongjie Zhang and Prof. Yi Wu, and earned his B.S. from the University of Science and Technology of China (USTC) in 2020.