Oct 9, 2021 · This work studies the question of Representation Learning in RL: how can we learn a compact low-dimensional representation such that on top of the ...
In addition to the online exploration setting, we also show that our new techniques can be directly used for designing offline RL algorithms for low-rank MDPs ...
Jan 6, 2022 · We develop new representation learning RL algorithms that enable sample efficient learning in low-rank MDPs under both online and offline ...
We propose an efficient algorithm, called ReLEX, for representation learning in both online and offline RL. Specifically, we show that the online version of ...
Oct 9, 2021 · Provably efficient representation selection in Low-rank Markov Decision Processes: from online to offline RL · Computer Science, Mathematics.
People also ask
What is online and offline reinforcement learning?
Is offline RL supervised learning?
What is unsupervised representation learning?
Aug 18, 2022 · Bibliographic details on Representation Learning for Online and Offline RL in Low-rank MDPs.
Improved online Representation Learning algorithm for low-rank MDP: Oracle-efficient + tight sample complexity. 2. New offline RL algorithm for low-rank MDP:.
Abstract. The low-rank MDP has emerged as an important model for studying representation learning and exploration in reinforcement learning.
Feb 1, 2023 · In this paper we focus on reward-free RL under low-rank MDP models, in which both the representation and linear weight vectors are unknown.
We propose an effi- cient algorithm, called ReLEX, for representation learning in both online and offline RL. Specifically, we show that the online version of ...