Pengxiang Li

I am a first-year PhD student in Beijing Institute of Technology(BIT), advised by Dr. Yuwei Wu and Dr. Yunde Jia. I am also a member of the joint PhD program ('TONG Program') with Beijing Institute for General Artificial Intelligence(BIGAI), and I am grateful to be advised by Dr. Qing Li and Dr. Zhi Gao. Previously, I got my Bachelor's degree in Computer Science and Technology from BIT in 2021.

My research interests lie in Vision and Language, non-Euclidean representation learning, and 3D vision. Specifically, I am interested in building the feedback refining systems for multi-modal models.

Email / Github

News

[2024.09] 🌟 One journal paper on Stereo Matching is accepted by T-CSVT.

Research

	Task-oriented Sequential Grounding in 3D Scenes Zhuofan Zhang, Ziyu Zhu, Pengxiang Li, Tengyu Liu, Xiaojian Ma, Yixin Chen, Baoxiong Jia, Siyuan Huang, Qing Li Preprint, 2024 [Arxiv] [Website] [Code] [Dataset] [Demo] [YouTube] We proposed a new task, Task-oriented Sequential Grounding in 3D scenes, and introduced SG3D, a large-scale dataset with 22,346 tasks and 112,236 steps in 4,895 real-world 3D scenes.
	FIRE: A Dataset for Feedback Integration and Refinement Evaluation of Multimodal Models Pengxiang Li, Zhi Gao, Bofei Zhang, Tao Yuan, Yuwei Wu, Mehrtash Harandi, Yunde Jia, Song-Chun Zhu, Qing Li Preprint, 2024 [Arxiv] [Website] [Code] [Dataset] [Model] [YouTube] A feedback-refinement dataset with 1.1M multi-turn conversations, which empowers VLMs to refine their responses based on given feedback.*
	Inter-Scale Similarity Guided Cost Aggregation for Stereo Matching Pengxiang Li, Chengtang Yao, Yunde Jia, Yuwei Wu Early Accepted by T-CSVT, 2024 [Paper] A plug-and-play module of inter-scale similarity guided cost aggregation to adaptively recover details in fine-grained areas for stereo matching.
	Hyperbolic Learning: Theory and Applications Pengxiang Li, Peilin Yu, Yangkai Xue, Yuwei Wu , Zhi Gao Tutorial, 2023 [Slide] A tutorial explores hyperbolic learning's theoretical underpinnings and applications, highlighting its advantages in modeling hierarchical data in diverse downstream felds.
	A Decomposition Model for Stereo Matching Chengtang Yao, Yunde Jia, Huijun Di* , Pengxiang Li, Yuwei Wu CVPR, 2021 [Paper] [Code] [Supp] A a decomposition model for stereo matching to solve the problem of excessive growth in computational cost (time and memory cost) as the resolution increases.

Experience

	Beijing Institute for General Artificial Intelligence(BIGAI), China 2024.02 - Now Joint training PhD student Advisor: Dr. Qing Li and Dr. Zhi Gao.
	Beijing Institute of Technology, China Master student 2021.09 - 2023.07 PhD student 2023.9 - Now Advisor: Dr. Yuwei Wu and Dr. Yunde Jia
	Beijing Institute of Technology, China 2017.08 - 2021.06 Undergraduate Student Advisor: Dr. Xian-Ling Mao

Thanks for your visiting by 😊.
Template stolen from Jon Barron. Page icon made by Freepik from flaticon.com .