ToyBox: Better Atari Environments for Testing Reinforcement Learning Agents

By John Foley, Emma Tosch, Kaleigh Clary, David D. Jensen. 2018

7 citations recorded by the source corpus.

It is a widely accepted principle that software without tests has bugs. Testing reinforcement learning agents is especially difficult because of the stochastic nature of both agents and environments, the complexity of state-of-the-art models, and the sequential nature of their predictions. Recently, the Arcade Learning Environment (ALE) has become one of the most widely used benchmark suites for deep learning research, and state-of-the-art Reinforcement Learning (RL) agents have been shown to routinely equal or exceed human performance on many ALE tasks. Since ALE is based on emulation of original Atari games, the environment does not provide semantically meaningful representations of internal game state. This means that ALE has limited utility as an environment for supporting testing or model introspection. We propose ToyBox, a collection of reimplementations of these games that solves this critical problem and enables robust testing of RL agents.

Subfields: Artificial Intelligence, Management Science and Operations Research

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Akanksha Atrey
2. Kaleigh Clary
3. David D. Jensen

In Management Science and Operations Research

1. David D. Jensen
2. Kaleigh Clary

Citations in the corpus

• Exploratory Not Explanatory: Counterfactual Analysis of Saliency Maps for Deep Reinforcement Learning (2019, methodology): “We employ T OY B OX , a set of fully parameterized implementation of Atari games (Foley et al., 2018), to generate interventional data under counterfactual conditions.”