TREND: Tri-teaching for Robust Preference-based Reinforcement Learning with Demonstrations
Shuaiyi Huang, Mara Levy, Anubhav Gupta, Daniel Ekpo, Ruijie Zheng, Abhinav Shrivastava
- Year
- 2025
- Access
- Open access
Abstract
Preference feedback collected by human or VLM annotators is often noisy, presenting a significant challenge for preference-based reinforcement learning that relies on accurate preference labels. To address this challenge, we propose TREND, a novel framework that integrates few-shot expert demonstrations with a tri-teaching strategy for effective noise mitigation. Our method trains three reward models simultaneously, where each model views its small-loss preference pairs as useful knowledge and teaches such useful pairs to its peer network for updating the parameters. Remarkably, our approach requires as few as one to three expert demonstrations to achieve high performance. We evaluate TREND on various robotic manipulation tasks, achieving up to 90% success rates even with noise levels as high as 40%, highlighting its effective robustness in handling noisy preference feedback. Project page: https://shuaiyihuang.github.io/publications/TREND.
Keywords
Related papers
State-of-the-art in mobile robot-assisted grinding technologies for large-scale complex components
Yusen Li, Ziwei Wang, Xiangye Zhu +9 more
Robotics and Computer-Integrated Manufacturing · 2026
A fusion prediction model of tool wear based on physical information and machine learning in five-axis milling TC4 titanium alloy
Shaoqing Qin, Lida Zhu, Yanpeng Hao +7 more
Robotics and Computer-Integrated Manufacturing · 2026
Enhancing robotic milling quality via a novel piezoelectric active damping toolholder
Bo Li, Yuanbo Zhao, Huijie Xiao +3 more
Robotics and Computer-Integrated Manufacturing · 2026
A novel method of suppressing low-frequency chatter in robotic milling using magnetically-induced nonlinear broadband multidirectional passive vibration absorber
Hao Li, Yuhui Yu, Rui Fu +3 more
Robotics and Computer-Integrated Manufacturing · 2026