On Inductive Biases in Deep Reinforcement Learning

TL;DR

This paper examines the role of inductive biases in deep reinforcement learning, analyzing how domain-specific components versus adaptive solutions affect performance and generality across various tasks.

Contribution

It provides an empirical comparison of domain-specific and adaptive components in deep RL, highlighting the trade-offs between performance and generality.

Findings

Adaptive components sometimes outperform domain-specific ones.

Fewer domain-specific components can enhance transfer to new tasks.

Performance varies depending on the environment and component adaptation.

Abstract

Many deep reinforcement learning algorithms contain inductive biases that sculpt the agent's objective and its interface to the environment. These inductive biases can take many forms, including domain knowledge and pretuned hyper-parameters. In general, there is a trade-off between generality and performance when algorithms use such biases. Stronger biases can lead to faster learning, but weaker biases can potentially lead to more general algorithms. This trade-off is important because inductive biases are not free; substantial effort may be required to obtain relevant domain knowledge or to tune hyper-parameters effectively. In this paper, we re-examine several domain-specific components that bias the objective and the environmental interface of common deep reinforcement learning agents. We investigated whether the performance deteriorates when these components are replaced with adaptive solutions from the literature. In our experiments, performance sometimes decreased with the adaptive components, as one might expect when comparing to components crafted for the domain, but sometimes the adaptive components performed better. We investigated the main benefit of having fewer domain-specific components, by comparing the learning performance of the two systems on a different set of continuous control problems, without additional tuning of either system. As hypothesized, the system with adaptive components performed better on many of the new tasks.