Learning to Query Internet Text for Informing Reinforcement Learning Agents

TL;DR

This paper explores training reinforcement learning agents to effectively query natural language sources like forums and wikis, improving out-of-distribution generalization by learning when and how to retrieve useful information.

Contribution

It introduces a method for training RL agents to query natural language sources using pretrained QA models and learning query timing, addressing challenges of noisy, large, real-world data.

Findings

Pretrained QA models excel at zero-shot querying in the target domain.

Agents learn to query at optimal times to maximize reward.

The approach improves generalization in reinforcement learning tasks.

Abstract

Generalization to out of distribution tasks in reinforcement learning is a challenging problem. One successful approach improves generalization by conditioning policies on task or environment descriptions that provide information about the current transition or reward functions. Previously, these descriptions were often expressed as generated or crowd sourced text. In this work, we begin to tackle the problem of extracting useful information from natural language found in the wild (e.g. internet forums, documentation, and wikis). These natural, pre-existing sources are especially challenging, noisy, and large and present novel challenges compared to previous approaches. We propose to address these challenges by training reinforcement learning agents to learn to query these sources as a human would, and we experiment with how and when an agent should query. To address the \textit{how}, we demonstrate that pretrained QA models perform well at executing zero-shot queries in our target domain. Using information retrieved by a QA model, we train an agent to learn \textit{when} it should execute queries. We show that our method correctly learns to execute queries to maximize reward in a reinforcement learning setting.