PIGDreamer: Privileged Information Guided World Models for Safe Partially Observable Reinforcement Learning
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915507779665920 |
|---|---|
| author | Huang, Dongchi Wang, Jiaqi Li, Yang Xia, Chunhe Zhang, Tianle Zhang, Kaige |
| author_facet | Huang, Dongchi Wang, Jiaqi Li, Yang Xia, Chunhe Zhang, Tianle Zhang, Kaige |
| contents | Partial observability presents a significant challenge for Safe Reinforcement Learning (Safe RL), as it impedes the identification of potential risks and rewards. Leveraging specific types of privileged information during training to mitigate the effects of partial observability has yielded notable empirical successes. In this paper, we propose Asymmetric Constrained Partially Observable Markov Decision Processes (ACPOMDPs) to theoretically examine the advantages of incorporating privileged information in Safe RL. Building upon ACPOMDPs, we propose the Privileged Information Guided Dreamer (PIGDreamer), a model-based RL approach that leverages privileged information to enhance the agent's safety and performance through privileged representation alignment and an asymmetric actor-critic structure. Our empirical results demonstrate that PIGDreamer significantly outperforms existing Safe RL methods. Furthermore, compared to alternative privileged RL methods, our approach exhibits enhanced performance, robustness, and efficiency. Codes are available at: https://github.com/hggforget/PIGDreamer. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_02159 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | PIGDreamer: Privileged Information Guided World Models for Safe Partially Observable Reinforcement Learning Huang, Dongchi Wang, Jiaqi Li, Yang Xia, Chunhe Zhang, Tianle Zhang, Kaige Machine Learning Partial observability presents a significant challenge for Safe Reinforcement Learning (Safe RL), as it impedes the identification of potential risks and rewards. Leveraging specific types of privileged information during training to mitigate the effects of partial observability has yielded notable empirical successes. In this paper, we propose Asymmetric Constrained Partially Observable Markov Decision Processes (ACPOMDPs) to theoretically examine the advantages of incorporating privileged information in Safe RL. Building upon ACPOMDPs, we propose the Privileged Information Guided Dreamer (PIGDreamer), a model-based RL approach that leverages privileged information to enhance the agent's safety and performance through privileged representation alignment and an asymmetric actor-critic structure. Our empirical results demonstrate that PIGDreamer significantly outperforms existing Safe RL methods. Furthermore, compared to alternative privileged RL methods, our approach exhibits enhanced performance, robustness, and efficiency. Codes are available at: https://github.com/hggforget/PIGDreamer. |
| title | PIGDreamer: Privileged Information Guided World Models for Safe Partially Observable Reinforcement Learning |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2508.02159 |