Xiangtong Yao


Picture of Xiangtong Yao

Xiangtong Yao

Technical University of Munich

Informatics 6 - Chair of Robotics, Artificial Intelligence and Real-time Systems (Prof. Knoll)

Place of employment

Informatics 6 - Chair of Robotics, Artificial Intelligence and Real-time Systems (Prof. Knoll)

Work:
Parkring 13, Room 2.02.20
85748 Garching b. München

Curriculum Vitae

Xiangtong Yao is currently a Ph.D. student in the Chair of Robotics, AI and Real-time Systems. His research advisor is Prof. Alois Knoll. He received his M.Eng degree in School of Computer Science and Engineering at Sun Yat-sen University, China.  Before he come to TUM to pursue his doctoral degree, he worked at Sun Yat-sen University as a research assistant. His research interests include Language-conditioned Meta-Reinforcement Learning for Manipulation tasks and Human-robot Interaction with New Tactile Sensor, such as triboelectric nanogenerator (TENG).

Thesis Topic

Topic 1: Language-conditioned Meta-Reinforcement Learning for Manipaulation Tasks

Meta-reinforcement learning (meta-RL) is a promising approach that enables the agent to learn new tasks quickly. However, most meta-RL algorithms show poor generalization in multiple-task scenarios due to the insufficient task information provided only by rewards. Inspired by teaching processing, Language-conditioned Meta-RL improves the generalization by matching language instructions and the agent’s behaviors. In this topic, our goal is combining NLP technologies with Meta-RL for manipulation tasks to enable robot to solve multiple complex manipulation tasks. For example, combining synonyms and antonyms of language instructions with robot's behaviors can improve the learning efficiency and generalization of robot in solving symmetrical tasks. For more information, please visit [2209.10656v1] Learning from Symmetry: Meta-Reinforcement Learning with Symmetric Data and Language Instructions (arxiv.org)

1: Language as abstractions for Meta-reinforcement learning

2: Learning from Symmetry: Meta-Reinforcement Learning with Symmetric Data and Language Instructions

If you are interested in the above topics, please feel free to contact me indicating your background and skills.

Demo

Publications

2023

  • Wang, Mingyang; Bing, Zhenshan; Yao, Xiangtong; Wang, Shuai; Huang, Kai; Su, Hang; Yang, Chenguang; Knoll, Alois: Meta-Reinforcement Learning Based on Self-Supervised Task Representation Learning. Association for the Advancement of Artificial Intelligence, AAAI, 2023 more…

2021

  • Shan, Yunxiao; Yao, Xiangtong; Lin, Hongquan; Zou, Xuesong; Huang, Kai: Lidar-Based Stable Navigable Region Detection for Unmanned Surface Vehicles. IEEE Transactions on Instrumentation and Measurement 70, 2021, 1-13 more…

2019

  • Chen, Longsheng; Chen, Yuanpeng; Yao, Xiangtong; Shan, Yunxiao; Chen, Long: An Adaptive Path Tracking Controller Based on Reinforcement Learning with Urban Driving Application. 2019 IEEE Intelligent Vehicles Symposium (IV), IEEE, 2019 more…
  • Yao, Xiangtong; Shan, Yunxiao; Li, Jieling; Ma, Donghui; Huang, Kai: LiDAR Based Navigable Region Detection for Unmanned Surface Vehicles. 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, 2019 more…