Learning from the memory of Atari 2600

TL;DR

This paper explores neural network agents trained on Atari 2600 games using only RAM data, achieving comparable or better performance than screen-based models, with some RAM-only agents outperforming benchmarks.

Contribution

It demonstrates that RAM data alone can be sufficient for effective game-playing agents, challenging the reliance on visual input.

Findings

RAM-only agents outperform screen-only in Seaquest.

Mixing screen and RAM did not improve performance.

Comparable results to benchmark models across games.

Abstract

We train a number of neural networks to play games Bowling, Breakout and Seaquest using information stored in the memory of a video game console Atari 2600. We consider four models of neural networks which differ in size and architecture: two networks which use only information contained in the RAM and two mixed networks which use both information in the RAM and information from the screen. As the benchmark we used the convolutional model proposed in NIPS and received comparable results in all considered games. Quite surprisingly, in the case of Seaquest we were able to train RAM-only agents which behave better than the benchmark screen-only agent. Mixing screen and RAM did not lead to an improved performance comparing to screen-only and RAM-only agents.