Adaptive Action Duration with Contextual Bandits for Deep Reinforcement Learning in Dynamic Environments

TL;DR

This paper introduces a method combining contextual bandits with deep reinforcement learning to adaptively select action durations, improving performance and efficiency in dynamic environments like Atari games.

Contribution

It presents a novel framework that integrates contextual bandits with DRL to adaptively determine action durations, enhancing policy flexibility and computational efficiency.

Findings

Significant performance improvements over static duration baselines.

Effective adaptation of action durations in Atari 2600 games.

Enhanced scalability for real-time applications like robotics.

Abstract

Deep Reinforcement Learning (DRL) has achieved remarkable success in complex sequential decision-making tasks, such as playing Atari 2600 games and mastering board games. A critical yet underexplored aspect of DRL is the temporal scale of action execution. We propose a novel paradigm that integrates contextual bandits with DRL to adaptively select action durations, enhancing policy flexibility and computational efficiency. Our approach augments a Deep Q-Network (DQN) with a contextual bandit module that learns to choose optimal action repetition rates based on state contexts. Experiments on Atari 2600 games demonstrate significant performance improvements over static duration baselines, highlighting the efficacy of adaptive temporal abstractions in DRL. This paradigm offers a scalable solution for real-time applications like gaming and robotics, where dynamic action durations are critical.