Tao Pu (蒲韬)

I am currently an R&D Director at X-Era Lab, leading foundational research on embodied AI. I received my Ph.D. degree from Sun Yat-sen University in 2025 under the supervision of Prof. Liang Lin, with joint doctoral training at Peking University. Prior to that, I received my Bachelor’s degree from Sun Yat-sen University in 2020.

North-star: Creating general-purpose robots that serve human life, physically and mentally, in anywhere.

Email  |  GitHub  |  Google Scholar  |  Zhihu

headshot
News
  • 12/2025 Promoted to R&D Director at X-Era Lab, leading research on embodied foundation models and building full-stack infrastructure for embodied AI systems.
  • 06/2024 Join X-Era Lab as a Senior Researcher Intern, focusing on world models and large-scale pretraining for embodied foundation models.
  • 09/2023 Join PKU CoRe Lab under the supervision of Prof. Yixin Zhu, focusing on visual cognition and humanoid robotics (i.e., dexterous manipulation and whole-body control).
---- show more ----
Selected Publications
Please see this page for more recent works and arXiv papers.
* denotes equal contribution, † denotes corresponding author.
Spatial–Temporal Knowledge-Embedded Transformer for Video Scene Graph Generation
Tao Pu*, Tianshui Chen* ✉, Hefeng Wu, Yongyi Lu, Liang Lin
IEEE Transactions on Image Processing (TIP), 2024
Paper | Code | Cite

TL;DR: A video scene graph generation approach that incorporates the prior spatial-temporal knowledge to learn better representations.

Heterogeneous Semantic Transfer for Multi-label Recognition with Partial Labels
Tianshui Chen* ✉, Tao Pu*, Lingbo Liu, Yukai Shi, Zhijing Yang, Liang Lin
International Journal of Computer Vision (IJCV), 2024
Paper | Code | Cite | 知乎

TL;DR: A weakly-supervised MLR approach that explores heterogeneous semantics inherently within multi-label images.

Cross-Domain Facial Expression Recognition: A Unified Evaluation Benchmark and Adversarial Graph Learning
Tianshui Chen*, Tao Pu*, Hefeng Wu ✉, Yuan Xie, Lingbo Liu, Liang Lin
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2022
Paper | Code | Cite | 知乎

TL;DR: One of the largest CD-FER evaluation benchmarks that unifies the source/target datasets and feature extractors for existing algorithms.

Honors & Awards
  • 中国科协青年人才托举工程博士生专项计划(首批), 2024
  • Graduate National Scholarship (博士研究生国家奖学金), 2024
  • Undergraduate National Scholarship (本科生国家奖学金), 2018
---- show more ----
Academic Services
  • Conference Reviewer: CVPR, ICCV, ICLR, NeurIPS, ICML, IJCAI, ACM MM, AAAI
  • Journal Reviewer: TPAMI, TNNLS, TKDE, TOMM, TCSVT