RIME: Robust Preference-based Reinforcement Learning with Noisy Preferences
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912089048612864 |
|---|---|
| author | Cheng, Jie Xiong, Gang Dai, Xingyuan Miao, Qinghai Lv, Yisheng Wang, Fei-Yue |
| author_facet | Cheng, Jie Xiong, Gang Dai, Xingyuan Miao, Qinghai Lv, Yisheng Wang, Fei-Yue |
| contents | Preference-based Reinforcement Learning (PbRL) circumvents the need for reward engineering by harnessing human preferences as the reward signal. However, current PbRL methods excessively depend on high-quality feedback from domain experts, which results in a lack of robustness. In this paper, we present RIME, a robust PbRL algorithm for effective reward learning from noisy preferences. Our method utilizes a sample selection-based discriminator to dynamically filter out noise and ensure robust training. To counteract the cumulative error stemming from incorrect selection, we suggest a warm start for the reward model, which additionally bridges the performance gap during the transition from pre-training to online training in PbRL. Our experiments on robotic manipulation and locomotion tasks demonstrate that RIME significantly enhances the robustness of the state-of-the-art PbRL method. Code is available at https://github.com/CJReinforce/RIME_ICML2024. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_17257 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | RIME: Robust Preference-based Reinforcement Learning with Noisy Preferences Cheng, Jie Xiong, Gang Dai, Xingyuan Miao, Qinghai Lv, Yisheng Wang, Fei-Yue Machine Learning Artificial Intelligence Robotics Preference-based Reinforcement Learning (PbRL) circumvents the need for reward engineering by harnessing human preferences as the reward signal. However, current PbRL methods excessively depend on high-quality feedback from domain experts, which results in a lack of robustness. In this paper, we present RIME, a robust PbRL algorithm for effective reward learning from noisy preferences. Our method utilizes a sample selection-based discriminator to dynamically filter out noise and ensure robust training. To counteract the cumulative error stemming from incorrect selection, we suggest a warm start for the reward model, which additionally bridges the performance gap during the transition from pre-training to online training in PbRL. Our experiments on robotic manipulation and locomotion tasks demonstrate that RIME significantly enhances the robustness of the state-of-the-art PbRL method. Code is available at https://github.com/CJReinforce/RIME_ICML2024. |
| title | RIME: Robust Preference-based Reinforcement Learning with Noisy Preferences |
| topic | Machine Learning Artificial Intelligence Robotics |
| url | https://arxiv.org/abs/2402.17257 |