A Bayesian Approach to Robust Inverse Reinforcement Learning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866910400827621376 |
|---|---|
| author | Wei, Ran Zeng, Siliang Li, Chenliang Garcia, Alfredo McDonald, Anthony Hong, Mingyi |
| author_facet | Wei, Ran Zeng, Siliang Li, Chenliang Garcia, Alfredo McDonald, Anthony Hong, Mingyi |
| contents | We consider a Bayesian approach to offline model-based inverse reinforcement learning (IRL). The proposed framework differs from existing offline model-based IRL approaches by performing simultaneous estimation of the expert's reward function and subjective model of environment dynamics. We make use of a class of prior distributions which parameterizes how accurate the expert's model of the environment is to develop efficient algorithms to estimate the expert's reward and subjective dynamics in high-dimensional settings. Our analysis reveals a novel insight that the estimated policy exhibits robust performance when the expert is believed (a priori) to have a highly accurate model of the environment. We verify this observation in the MuJoCo environments and show that our algorithms outperform state-of-the-art offline IRL algorithms. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2309_08571 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | A Bayesian Approach to Robust Inverse Reinforcement Learning Wei, Ran Zeng, Siliang Li, Chenliang Garcia, Alfredo McDonald, Anthony Hong, Mingyi Machine Learning We consider a Bayesian approach to offline model-based inverse reinforcement learning (IRL). The proposed framework differs from existing offline model-based IRL approaches by performing simultaneous estimation of the expert's reward function and subjective model of environment dynamics. We make use of a class of prior distributions which parameterizes how accurate the expert's model of the environment is to develop efficient algorithms to estimate the expert's reward and subjective dynamics in high-dimensional settings. Our analysis reveals a novel insight that the estimated policy exhibits robust performance when the expert is believed (a priori) to have a highly accurate model of the environment. We verify this observation in the MuJoCo environments and show that our algorithms outperform state-of-the-art offline IRL algorithms. |
| title | A Bayesian Approach to Robust Inverse Reinforcement Learning |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2309.08571 |