Dual Action Policy for Robust Sim-to-Real Reinforcement Learning
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_ | 1866914974501175296 |
|---|---|
| author | Terence, Ng Wen Zheng Jianda, Chen |
| author_facet | Terence, Ng Wen Zheng Jianda, Chen |
| contents | This paper presents Dual Action Policy (DAP), a novel approach to address the dynamics mismatch inherent in the sim-to-real gap of reinforcement learning. DAP uses a single policy to predict two sets of actions: one for maximizing task rewards in simulation and another specifically for domain adaptation via reward adjustments. This decoupling makes it easier to maximize the overall reward in the source domain during training. Additionally, DAP incorporates uncertainty-based exploration during training to enhance agent robustness. Experimental results demonstrate DAP's effectiveness in bridging the sim-to-real gap, outperforming baselines on challenging tasks in simulation, and further improvement is achieved by incorporating uncertainty estimation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_12250 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Dual Action Policy for Robust Sim-to-Real Reinforcement Learning Terence, Ng Wen Zheng Jianda, Chen Machine Learning Artificial Intelligence Robotics This paper presents Dual Action Policy (DAP), a novel approach to address the dynamics mismatch inherent in the sim-to-real gap of reinforcement learning. DAP uses a single policy to predict two sets of actions: one for maximizing task rewards in simulation and another specifically for domain adaptation via reward adjustments. This decoupling makes it easier to maximize the overall reward in the source domain during training. Additionally, DAP incorporates uncertainty-based exploration during training to enhance agent robustness. Experimental results demonstrate DAP's effectiveness in bridging the sim-to-real gap, outperforming baselines on challenging tasks in simulation, and further improvement is achieved by incorporating uncertainty estimation. |
| title | Dual Action Policy for Robust Sim-to-Real Reinforcement Learning |
| topic | Machine Learning Artificial Intelligence Robotics |
| url | https://arxiv.org/abs/2410.12250 |