Timezone: »

 
Poster
Near Optimal Reward-Free Reinforcement Learning
Zhang Zihan · Simon Du · Xiangyang Ji

Wed Jul 21 09:00 AM -- 11:00 AM (PDT) @ Virtual #None
in Poster Session 3 »
We study the reward-free reinforcement learning framework, which is particularly suitable for batch reinforcement learning and scenarios where one needs policies for multiple reward functions. This framework has two phases: in the exploration phase, the agent collects trajectories by interacting with the environment without using any reward signal; in the planning phase, the agent needs to return a near-optimal policy for arbitrary reward functions. %This framework is suitable for batch RL setting and the setting where there are multiple reward functions of interes We give a new efficient algorithm, \textbf{S}taged \textbf{S}ampling + \textbf{T}runcated \textbf{P}lanning (\algoname), which interacts with the environment at most $O\left( \frac{S^2A}{\epsilon^2}\poly\log\left(\frac{SAH}{\epsilon}\right) \right)$ episodes in the exploration phase, and guarantees to output a near-optimal policy for arbitrary reward functions in the planning phase, where $S$ is the size of state space, $A$ is the size of action space, $H$ is the planning horizon, and $\epsilon$ is the target accuracy relative to the total reward. Notably, our sample complexity scales only \emph{logarithmically} with $H$, in contrast to all existing results which scale \emph{polynomially} with $H$. Furthermore, this bound matches the minimax lower bound $\Omega\left(\frac{S^2A}{\epsilon^2}\right)$ up to logarithmic factors. Our results rely on three new techniques : 1) A new sufficient condition for the dataset to plan for an $\epsilon$-suboptimal policy % for any totally bounded reward function ; 2) A new way to plan efficiently under the proposed condition using soft-truncated planning; 3) Constructing extended MDP to maximize the truncated accumulative rewards efficiently.
Keywords: RL, Decisions and Control Theory

Author Information

Zhang Zihan (Tsinghua University)
Simon Du (University of Washington)
Xiangyang Ji (Tsinghua University)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors