Boosting Soft Q-Learning by Bounding
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916301612515328 |
|---|---|
| author | Adamczyk, Jacob Makarenko, Volodymyr Tiomkin, Stas Kulkarni, Rahul V. |
| author_facet | Adamczyk, Jacob Makarenko, Volodymyr Tiomkin, Stas Kulkarni, Rahul V. |
| contents | An agent's ability to leverage past experience is critical for efficiently solving new tasks. Prior work has focused on using value function estimates to obtain zero-shot approximations for solutions to a new task. In soft Q-learning, we show how any value function estimate can also be used to derive double-sided bounds on the optimal value function. The derived bounds lead to new approaches for boosting training performance which we validate experimentally. Notably, we find that the proposed framework suggests an alternative method for updating the Q-function, leading to boosted performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_18033 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Boosting Soft Q-Learning by Bounding Adamczyk, Jacob Makarenko, Volodymyr Tiomkin, Stas Kulkarni, Rahul V. Machine Learning Artificial Intelligence An agent's ability to leverage past experience is critical for efficiently solving new tasks. Prior work has focused on using value function estimates to obtain zero-shot approximations for solutions to a new task. In soft Q-learning, we show how any value function estimate can also be used to derive double-sided bounds on the optimal value function. The derived bounds lead to new approaches for boosting training performance which we validate experimentally. Notably, we find that the proposed framework suggests an alternative method for updating the Q-function, leading to boosted performance. |
| title | Boosting Soft Q-Learning by Bounding |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2406.18033 |