Shusheng Yang
NYU Courant PhD student; first author of Cambrian-S and co-first author of Thinking in Space and Qwen-VL
- Ph.D. student at NYU Courant, advised by Prof. Saining Xie, New York University (Courant Institute)1
Shusheng Yang is a PhD student at NYU Courant advised by Saining Xie, working where computer vision meets multimodal learning, with a focus on representation learning, world modeling, spatial intelligence and long-context modeling. Yang is first author of Cambrian-S (2025), which introduces the VSI-SUPER benchmark and the VSI-590K training data for spatial supersensing in video, and co-first author of Thinking in Space (CVPR 2025), which introduced the VSI-Bench visual-spatial benchmark. Yang also co-authored Cambrian-1, a vision-centric family of multimodal language models released with weights, code and data, and was a co-first author of the Qwen-VL technical report.12468
Work
Cambrian-S (external site: github.com)Open-source code
First author (equal contribution with Jihan Yang) of "Cambrian-S: Towards Spatial Supersensing in Video" (2025; ICLR 2026), which presents the VSI-SUPER benchmark, curates the VSI-590K dataset and trains Cambrian-S models; the code repository is licensed under Apache 2.0.213
Thinking in Space (VSI-Bench) (external site: github.com)Open-source code
Co-first author of "Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces" (CVPR 2025 oral), which introduced VSI-Bench, a video-based visual-spatial intelligence benchmark; the code repository is licensed under Apache 2.0.415
Cambrian-1 (external site: github.com)Open weights
Co-author of "Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs" (NeurIPS 2024 oral), whose authors provide model weights, code, supporting tools and datasets; the code repository is licensed under Apache 2.0.617
Qwen-VL (external site: github.com)Open weights
Co-first author with Jinze Bai and Shuai Bai of the Qwen-VL technical report (2023), which introduces the Qwen-VL and Qwen-VL-Chat vision-language models and says code, demo and models are available; the repository's license file is the Tongyi Qianwen license agreement.819
This profile covers public professional work only, from the sources below. It is not an endorsement or a ranking. To correct or remove it, use the correction form.