Zeyue Xue

Zeyue Xue

Research Scientist in Generative AI

Biography

I’m Zeyue Xue (薛泽岳), a Research Scientist at Joy Future Academy, JD.com, where I build generative AI products and platforms. I received my Ph.D. from the University of Hong Kong (MMLab@HKU) in 2026 under the supervision of Ping Luo and Wenping Wang.

I launched the Echo Project at Joy Future Academy and led the development and release of projects including JoyAI-Echo, JoyAI-Echo-1.5, and Echo-WM. My research focuses on large-scale generative models and interactive world models for games, real-world environments, and robotics.


Research Interests

  • Deep Generative Model
  • World Model

Representative Works

  1. Echo Team. “JoyAI-Echo”&”JoyAI-Echo-1.5”&”Echo-WM”, Code.
  2. Zeyue Xue, Jie Wu, Yu Gao, Fangyuan Kong, Lingting Zhu, Mengzhao Chen, Zhiheng Liu, Wei Liu, Qiushan Guo, Weilin Huang, Ping Luo. “DanceGRPO: Unleashing GRPO on Visual Generation”, Seed Tech Report, Code.
  3. Zeyue Xue*, Guanglu Song*, Qiushan Guo, Boxiao Liu, Zhuofan Zong, Yu Liu, Ping Luo. “RAPHAEL: Text-to-Image Generation via Large Mixture of Diffusion Paths”, NeurIPS 2023, Model API.
  4. Zeyue Xue, Jianming Liang, Guanglu Song, Zhuofan Zong, Liang Chen, Yu Liu, Ping Luo. “Large-batch Optimization for Dense Visual Predictions: Training Faster R-CNN in 4.2 minutes”, NeurIPS 2022, Code.

Invited Talks

2025-06: invited talks at PKU, ZJU, Xiaomi, Pika, Adobe, Anuttacon, ByteDance Seed about DanceGRPO. The example slide can be found here.