Weijia Mao

Biography

I am a fourth-year Ph.D. candidate in Show Lab @ NUS, advised by Prof. Mike Shou.

Selected topics and representative projects include:

I was fortunate to intern at TikTok, where I worked with Hao Chen and Zhenheng Yang.

I am open to discussions and collaborations, and actively seeking opportunities in the industry. Feel free to reach out.

2026 Feb: Adv-GRPO was accepted by CVPR 2026.
2025 Nov: Released Adv-GRPO , the first work to introduce adversarial training into RL for image quality optimization, leveraging high-quality reference images and a visual foundation model as the reward signal.
2025 May: Released UniRL, the first work to apply GRPO to unified multimodal models.
2025 Mar: Released our long-context video generation model FAR.
2025 Jan: Released UniMod, proposing an efficient training framework for unified models.
2024 Sep: Released Show-o, the first unified model that integrates both diffusion and autoregressive (AR) modeling.
2024 Aug: EgoExo-V was accepted by NeurIPS 2024.
2024 May: Released EgoExo-V, converting third-person videos into egocentric views using NeRF and video diffusion models.
2023 Nov: Released ShowRoom3D, the first work that combines NeRF and diffusion models for 3D scene generation.
2023 Jan: Joined Show Lab @ NUS to start my Ph.D. journey.
2022 Jan: Joined Show Lab @ NUS to start my Master’s journey.

	Show-o: A Unified Multimodal Model Integrating Diffusion and Autoregressive Modeling Jinheng Xie^†, Weijia Mao^†, Zechen Bai^†, David Junhao Zhang^†, Weihao Wang, Kevin Qinghong Lin, Yuchao Gu, Zhijie Chen, Zhenheng Yang, Mike Zheng Shou^. ICLR,* 2025. [project page] [paper] [code]
	Adv-GRPO: Adversarial Reinforcement Learning for Image Quality Optimization Weijia Mao, Hao Chen^, Zhenheng Yang, Mike Zheng Shou^. CVPR, 2026. [project page] [paper] [code]
	BitDance: Scaling Autoregressive Generative Models with Binary Tokens Yuang Ai, Jiaming Han, Shaobin Zhuang, Weijia Mao, Xuefeng Hu, Ziyan Yang, Zhenheng Yang, Huaibo Huang, Xiangyu Yue, Hao Chen. [project page] [paper] [code]
	EgoExo-V: Bridging Egocentric and Exocentric Video via NeRF and Diffusion Models Jia-Wei Liu^†, Weijia Mao^†, Zhongcong Xu, Jussi Keppo, Mike Zheng Shou. NeurIPS, 2024. [code] ^† Equal contribution.
	Long-Context Autoregressive Video Modeling with Next-Frame Prediction Yuchao Gu, Weijia Mao, Mike Zheng Shou. arXiv, 2025. [project page] [paper] [code]