Deep TAMER: Interactive Agent Shaping in High-Dimensional State Spaces

TL;DR

Deep TAMER integrates deep learning with human-in-the-loop reinforcement learning, enabling rapid training of agents in high-dimensional spaces using real-time feedback, demonstrated by outperforming humans in Atari Bowling with minimal feedback.

Contribution

It introduces Deep TAMER, a novel framework combining deep neural networks with human feedback for efficient learning in complex, high-dimensional environments.

Findings

Deep TAMER successfully trained an agent to outperform humans in Atari Bowling.

The framework requires only 15 minutes of human feedback for effective learning.

Deep TAMER outperforms traditional reinforcement learning methods in complex tasks.

Abstract

While recent advances in deep reinforcement learning have allowed autonomous learning agents to succeed at a variety of complex tasks, existing algorithms generally require a lot of training data. One way to increase the speed at which agents are able to learn to perform tasks is by leveraging the input of human trainers. Although such input can take many forms, real-time, scalar-valued feedback is especially useful in situations where it proves difficult or impossible for humans to provide expert demonstrations. Previous approaches have shown the usefulness of human input provided in this fashion (e.g., the TAMER framework), but they have thus far not considered high-dimensional state spaces or employed the use of deep learning. In this paper, we do both: we propose Deep TAMER, an extension of the TAMER framework that leverages the representational power of deep neural networks in order to learn complex tasks in just a short amount of time with a human trainer. We demonstrate Deep TAMER's success by using it and just 15 minutes of human-provided feedback to train an agent that performs better than humans on the Atari game of Bowling - a task that has proven difficult for even state-of-the-art reinforcement learning methods.