“Develop excellence with Westlake University.”
Biography
Huang Xiangru, born in 1991, participated in the National Olympiad in Informatics (NOI) in high school and was recommended for admission to the ACM pilot class at Shanghai JiaoTong University. After completing his undergraduate studies, he went to the University of Texas at Austin in the United States to pursue a PhD in computer science with a research focus on three-dimensional data processing, earning his doctorate in 2020. He then joined the Computer Science and Artificial Intelligence Laboratory (CSAIL) at the Massachusetts Institute of Technology, dedicating his efforts to the processing of large 3D data, 3D perception, and the development of 3D artificial intelligence generated content (AIGC) algorithms. He will subsequently serve as an assistant professor at Westlake University full-time, continuing his research in the relevant fields.
History
2021
Post-doctoral at CSAIL, Massachusetts Institute of Technology, working with Professor Justin Solomon
2014
Ph.D. in University of Texas at Austin, working with Professor Qixing Huang
2013
Research Assistant in Nanyang Technological University
2009
Shanghai JiaoTong University, ACM honored class
2008
NOI 2008, silver medal
Research
Professor Huang has made significant achievements in related fields in 3D artificial intelligence, including 1) Scanning and acquisition of 3D data, 2) 3D perception, 3) 3D generation.
1. 3D Scanning and Reconstruction Algorithms: He has completed a total of 8 papers in this direction, including a) systematically analyzing the reliability of traditional optimization algorithms in 3D scanning and reconstruction, proposing a series of algorithms based on prior knowledge; b) proposing the first machine learning-based multi-frame 3D reconstruction algorithm (Learn2Sync); c) proposing an automated 3D scanning scheme based on uncertainty estimation (UQ), which provides the possibility for automated 3D data acquisition.
2. 3D Perception: He has completed a total of 4 papers in this direction, including a) improving the performance of 3D perception algorithms by using geometric prior knowledge; b) improving algorithm performance and significantly enhancing model computational efficiency by using hybrid data representation.
3. 3D Generation: He has completed a total of 3 papers in this direction, including a) a graphics matching algorithm based on deep learning; b) a 3D human body generation and understanding algorithm based on physics and large 3D data.
Representative Publications
1. Xiangru Huang, Zhenxiao Liang, Xiaowei Zhou, Yao Xie, Leonidas Guibas, Qixing Huang*, Learning Transformation Synchronization, CVPR 2019
2. Qixing Huang+, Xiangru Huang+, Bo Sun+, Zaiwei Zhang, Junfeng Jiang, Chandrajit Bajaj* (+equal contribution), ARAPReg: An As-Rigid-As Possible Regularization Loss for Learning Deformable Shape Generators, International Conference on Computer Vision (ICCV) 2021.
2. Xiangru Huang+, Zhenxiao Liang+, Qixing Huang* (+equal contribution), Uncertainty Quantification for Multi-scan Registration, ACM Transactions on Graphics, 39(4), Proceedings of ACM SIGGRAPH 2020.
3. Xiangru Huang, Yue Wang, Vitor Guizilini, Rares Ambrus, Adrien Gaidon, Justin Solomon*, Representation Learning for Object Detection from Unlabeled Point Cloud Sequences, Conference on Robotic Learning (CoRL) 2022
4. Xiangru Huang, Haitao Yang, Etienne Vouga, Qixing Huang*, Dense Human Correspondence via Learning Transformation Synchronization on Graphs, 2020 Conference on Neural Information Processing Systems (NeurIPS 2020)
5. Xiangru Huang+, Zhenxiao Liang+, Chandrajit Bajaj, Qixing Huang (+equal contribution), Translation Synchronization via Truncated Least Squares, In Advances in Neural Information Processing Systems (NIPS), 2017
6. Haitao Yang, Xiangru Huang, Bo Sun, Chandrajit Bajaj, Qixing Huang*, GenCorres: Consistent Shape Matching via Coupled Implicit-Explicit Shape Generative Models, arXiv.
7. Ian E.H. Yen+, Xiangru Huang+, Kai Zhong, Pradeep Ravikumar, Inderjit S. Dhillon* (+ equal contribution), PD-Sparse: A Primal and Dual Sparse Approach to Extreme Multiclass and Multilabel Classification, In International Conference on Machine Learning (ICML), 2016.
8. Ian E.H. Yen, Xiangru Huang, Wei Dai, Pradeep Ravikumar, Inderjit S. Dhillon, Eric P. Xing*, PPDSparse: A Parallel Primal and Dual Sparse Method to Extreme Classification, ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2017.
Contact Us
Email: huangxiangru@westlake.edu.cn
We have several open positions for postdocs, graduate students and research assistants. Our lab develops fundamental and efficient algorithms and infrastructures, investing into the future of the 3D big data era. We target applications in computer graphics, 3D computer vision and 3D AIGC.
Applicants who are interested in our lab are encouraged to send emails and ask for a discussion.