Maximum Reward Formulation In Reinforcement Learning

Abstract

Reinforcement learning (RL) algorithms typically deal with maximizing the expected cumulative return (discounted or undiscounted, finite or infinite horizon). However, several crucial applications in the real world, such as drug discovery, do not fit within this framework because an RL agent only needs to identify states (molecules) that achieve the highest reward within a trajectory and does not need to optimize for the expected cumulative return. In this work, we formulate an objective function to maximize the expected maximum reward along a trajectory, derive a novel functional form of the Bellman equation, introduce the corresponding Bellman operators, and provide a proof of convergence. Using this formulation, we achieve state-of-the-art results on the task of molecule generation that mimics a real-world drug discovery pipeline.

arXiv comment14 pages, 5 figures Update based on reviewer feedback

References in corpus (6)

• Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
• Addressing Function Approximation Error in Actor-Critic Methods
• Junction Tree Variational Autoencoder for Molecular Graph Generation
• ChemBO: Bayesian Optimization of Small Organic Molecules with Synthesizable Recommendations
• Active Neural Localization
• Guiding Deep Molecular Optimization with Genetic Exploration

Cited by in corpus (1)

• Reinforcement Learning with Non-Cumulative Objective