Adaptive $Q$-Network: On-the-fly Target Selection for Deep Reinforcement Learning

TL;DR

AdaQN is a novel AutoRL method that adaptively selects hyperparameters on-the-fly without extra samples, improving efficiency, stability, and performance in deep RL tasks.

Contribution

Introduces AdaQN, a new adaptive hyperparameter selection method for RL that accounts for non-stationarity without additional samples, enhancing AutoRL applicability.

Findings

Improves sample-efficiency and training stability in RL tasks.

Demonstrates superior performance on MuJoCo and Atari benchmarks.

Shows robustness to stochastic environments.

Abstract

Deep Reinforcement Learning (RL) is well known for being highly sensitive to hyperparameters, requiring practitioners substantial efforts to optimize them for the problem at hand. This also limits the applicability of RL in real-world scenarios. In recent years, the field of automated Reinforcement Learning (AutoRL) has grown in popularity by trying to address this issue. However, these approaches typically hinge on additional samples to select well-performing hyperparameters, hindering sample-efficiency and practicality. Furthermore, most AutoRL methods are heavily based on already existing AutoML methods, which were originally developed neglecting the additional challenges inherent to RL due to its non-stationarities. In this work, we propose a new approach for AutoRL, called Adaptive $Q$-Network (AdaQN), that is tailored to RL to take into account the non-stationarity of the optimization procedure without requiring additional samples. AdaQN learns several $Q$-functions, each one trained with different hyperparameters, which are updated online using the $Q$-function with the smallest approximation error as a shared target. Our selection scheme simultaneously handles different hyperparameters while coping with the non-stationarity induced by the RL optimization procedure and being orthogonal to any critic-based RL algorithm. We demonstrate that AdaQN is theoretically sound and empirically validate it in MuJoCo control problems and Atari $2600$ games, showing benefits in sample-efficiency, overall performance, robustness to stochasticity and training stability.