Tell me why! Explanations support learning relational and causal structure

Inferring the abstract relational and causal structure of the world is a major challenge for reinforcement-learning (RL) agents. For humans, language—particularly in the form of explanations —plays a considerable role in overcoming this challenge. Here, we show that language can play a similar role for deep RL agents in complex environments. While agents typically struggle to acquire relational and causal knowledge, augmenting their experience by training them to predict language descriptions and explanations can overcome these limitations. We show that language can help agents learn challenging relational tasks, and examine which aspects of language contribute to its benefits. We then show that explanations can help agents to infer not only relational but also causal structure. Language can shape the way that agents to generalize out-of-distribution from ambiguous, causally-confounded training, and explanations even allow agents to learn to perform experimental interventions to identify causal relationships. Our results suggest that language description and explanation may be powerful tools for improving agent learning and generalization. It is often argued that machine learning models—and deep learning models in particular—lack the human proficiencies forming abstractions and inferring relational or causal These limitations can make it hard to train models that generalize out-of-distribution, or that reason in human-like ways, particularly for reinforcement learning (RL) agents that receive high-bandwidth input from raw pixels and must learn to act in partially-observable environments. the main text results without explanations show that these losses are not sufficient for learning either. This shows that the benefits of explanations are not simply due to having more supervision for the agent, but rather are specific to supervision that highlights the abstract task structure.