Computationally Efficient RL under Linear Bellman Completeness for Deterministic Dynamics
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_ | 1866913713818173440 |
|---|---|
| author | Wu, Runzhe Sekhari, Ayush Krishnamurthy, Akshay Sun, Wen |
| author_facet | Wu, Runzhe Sekhari, Ayush Krishnamurthy, Akshay Sun, Wen |
| contents | We study computationally and statistically efficient Reinforcement Learning algorithms for the linear Bellman Complete setting. This setting uses linear function approximation to capture value functions and unifies existing models like linear Markov Decision Processes (MDP) and Linear Quadratic Regulators (LQR). While it is known from the prior works that this setting is statistically tractable, it remained open whether a computationally efficient algorithm exists. Our work provides a computationally efficient algorithm for the linear Bellman complete setting that works for MDPs with large action spaces, random initial states, and random rewards but relies on the underlying dynamics to be deterministic. Our approach is based on randomization: we inject random noise into least squares regression problems to perform optimistic value iteration. Our key technical contribution is to carefully design the noise to only act in the null space of the training data to ensure optimism while circumventing a subtle error amplification issue. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_11810 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Computationally Efficient RL under Linear Bellman Completeness for Deterministic Dynamics Wu, Runzhe Sekhari, Ayush Krishnamurthy, Akshay Sun, Wen Machine Learning Robotics Systems and Control We study computationally and statistically efficient Reinforcement Learning algorithms for the linear Bellman Complete setting. This setting uses linear function approximation to capture value functions and unifies existing models like linear Markov Decision Processes (MDP) and Linear Quadratic Regulators (LQR). While it is known from the prior works that this setting is statistically tractable, it remained open whether a computationally efficient algorithm exists. Our work provides a computationally efficient algorithm for the linear Bellman complete setting that works for MDPs with large action spaces, random initial states, and random rewards but relies on the underlying dynamics to be deterministic. Our approach is based on randomization: we inject random noise into least squares regression problems to perform optimistic value iteration. Our key technical contribution is to carefully design the noise to only act in the null space of the training data to ensure optimism while circumventing a subtle error amplification issue. |
| title | Computationally Efficient RL under Linear Bellman Completeness for Deterministic Dynamics |
| topic | Machine Learning Robotics Systems and Control |
| url | https://arxiv.org/abs/2406.11810 |