Xuming He

Associate Professor PLUS Lab School of Information Science and Technology (SIST) ShanghaiTech University Address Room 1A-221, SIST Building 393 Middle Huaxia Road Pudong, Shanghai, 201210, China Contact Phone: +86 21 20685378 Email: hexm at shanghaitech dot edu dot cn or xmhe dot pluslab at google mail

Prospective students

• We are looking for prospective graduate students for 2025. You must contact me via email if you are interested in joining in my lab. 学科方向: 计算机科学与技术.

• We also have multiple research projects (minimal 6 months) for undergrad students in SIST.

• We are looking for a long-term research assistant (科研助理). Contact us via email if interested.

News

• [12/2024] One paper on open-vocabulary scene graph generation accepted in AAAI 2025.

• [11/2024] One paper on object grasping accepted in 3DV 2025.

• [09/2024] One paper on OOD semantic segmentation accepted in NeurIPS 2024.

• [07/2024] One paper on novel class discovery accepted in ECCV 2024.

Research Interest

• Open-world scene understanding

• Multi-modal learning and reasoning

• Continual learning and knowledge discovery

• Explainable artificial intelligence

Selected Recent Publications

• Generalize or Detect? Towards Robust Semantic Segmentation Under Multiple Distribution Shifts, [Arxiv]
  Zhitong Gao, Bingnan Li, Mathieu Salzmann, Xuming He
  Advances in Neural Information Processing Systems (NeurIPS), 2024

• Machine Learning on Microstructure–Property Relationship of Lithium-Ion Conducting Oxide Solid Electrolytes, [ACS Web]
  Yue Zhang, Xiaoyu Lin, Wenbo Zhai, Yanran Shen, Shaojie Chen, Yining Zhang, Yi Yu, Xuming He, Wei Liu
  Nano Lett., 2024, 24, 17, 5292–5300

• From Pixels to Graphs: Open-Vocabulary Scene Graph Generation with Vision-Language Models, [Arxiv]
  Rongjie Li*, Songyang Zhang*, Dahua Lin, Kai Chen, Xuming He
  IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024

• Learning by Correction: Efficient Tuning Task for Zero-Shot Generative Vision-Language Reasoning, [Arxiv]
  Rongjie Li*, Yu Wu*, Xuming He
  IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024

• SGTR+: End-to-end Scene Graph Generation with Transformer, [Arxiv]
  Rongjie Li, Songyang Zhang, Xuming He
  IEEE Transactions on Pattern Analysis and Machine Intelligence, Preprint, 2023

• Class-relation Knowledge Distillation for Novel Class Discovery, [Arxiv]
  Peiyan Gu, Chuyu Zhang, Ruiji Xu, Xuming He
  International Conference on Computer Vision (ICCV), 2023

• Modeling Multimodal Aleatoric Uncertainty in Segmentation with Mixture of Stochastic Experts, [Arxiv]
  Zhitong Gao, Yucong Chen, Chuyu Zhang, Xuming He
  International Conference on Learning Representation (ICLR), 2023

• Weakly-supervised HOI Detection via Prior-guided Bi-level Representation Learning, [Arxiv]
  Bo Wan, Yongfei Liu, Desen Zhou, Tinne Tuytelaars, Xuming He
  International Conference on Learning Representation (ICLR), 2023

• Distribution Alignment: A Unified Framework for Long-tail Visual Recognition, [Arxiv]
  Songyang Zhang, Zeming Li, Shipeng Yan, Xuming He, Jian Sun
  IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2021

• DER: Dynamically Expandable Representation for Class Incremental Learning, [Arxiv]
  Shipeng Yan, Jiangwei Xie, Xuming He
  IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2021