Can Agents Learn by Analogy? An Inferable Model for PAC Reinforcement Learning

Abstract

Model-based reinforcement learning algorithms make decisions by building and utilizing a model of the environment. However, none of the existing algorithms attempts to infer the dynamics of any state-action pair from known state-action pairs before meeting it for sufficient times. We propose a new model-based method called Greedy Inference Model (GIM) that infers the unknown dynamics from known dynamics based on the internal spectral properties of the environment. In other words, GIM can "learn by analogy". We further introduce a new exploration strategy which ensures that the agent rapidly and evenly visits unknown state-action pairs. GIM is much more computationally efficient than state-of-the-art model-based algorithms, as the number of dynamic programming operations is independent of the environment size. Lower sample complexity could also be achieved under mild conditions compared against methods without inferring. Experimental results demonstrate the effectiveness and efficiency of GIM in a variety of real-world tasks.

arXiv commentTo be published in proceedings of AAMAS 2020

References in corpus (9)

• Incoherence-Optimal Matrix Completion
• Contextual Decision Processes with Low Bellman Rank are PAC-Learnable
• Unifying PAC and Regret: Uniform PAC Bounds for Episodic Reinforcement Learning
• Reinforcement Learning in Feature Space: Matrix Bandit, Kernels, and Regret Bound
• Value function approximation via low-rank models
• Matrix Completion from $O(n)$ Samples in Linear Time
• Reinforcement Learning in Rich-Observation MDPs using Spectral Methods
• Estimation of Markov Chain via Rank-Constrained Likelihood
• Adaptive Low-Nonnegative-Rank Approximation for State Aggregation of Markov Chains