We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
Provably Efficient Reinforcement Learning with Linear Function Approximation.
- Authors
Jin, Chi; Yang, Zhuoran; Wang, Zhaoran; Jordan, Michael I.
- Abstract
Modern reinforcement learning (RL) is commonly applied to practical problems with an enormous number of states, where function approximation must be deployed to approximate either the value function or the policy. The introduction of function approximation raises a fundamental set of challenges involving computational and statistical efficiency, especially given the need to manage the exploration/exploitation trade-off. As a result, a core RL question remains open: how can we design provably efficient RL algorithms that incorporate function approximation? This question persists even in a basic setting with linear dynamics and linear rewards, for which only linear function approximation is needed. This paper presents the first provable RL algorithm with both polynomial run time and polynomial sample complexity in this linear setting, without requiring a "simulator" or additional assumptions. Concretely, we prove that an optimistic modification of least-squares value iteration—a classical algorithm frequently studied in the linear setting—achieves O ˜ ( d 3 H 3 T) regret, where d is the ambient dimension of feature space, H is the length of each episode, and T is the total number of steps. Importantly, such regret is independent of the number of states and actions. Funding: This work was supported by the Defense Advanced Research Projects Agency program on Lifelong Learning Machines.
- Subjects
UNITED States. Defense Advanced Research Projects Agency; REINFORCEMENT learning; MACHINE learning; POLYNOMIAL time algorithms; MILITARY research; ALGORITHMS
- Publication
Mathematics of Operations Research, 2023, Vol 48, Issue 3, p1496
- ISSN
0364-765X
- Publication type
Article
- DOI
10.1287/moor.2022.1309