Deep Apprenticeship Learning for Playing Games

TL;DR

This paper introduces a deep apprenticeship learning approach using expert behavior to train agents in complex games without explicit reward functions, demonstrating potential for future reinforcement learning advancements.

Contribution

It presents a novel apprenticeship learning method based on supervised learning techniques applied to deep neural networks for game playing tasks.

Findings

Method applied to Atari games with promising results

Shows potential for deep apprenticeship learning in complex tasks

Highlights future research directions in reward-free learning

Abstract

In the last decade, deep learning has achieved great success in machine learning tasks where the input data is represented with different levels of abstractions. Driven by the recent research in reinforcement learning using deep neural networks, we explore the feasibility of designing a learning model based on expert behaviour for complex, multidimensional tasks where reward function is not available. We propose a novel method for apprenticeship learning based on the previous research on supervised learning techniques in reinforcement learning. Our method is applied to video frames from Atari games in order to teach an artificial agent to play those games. Even though the reported results are not comparable with the state-of-the-art results in reinforcement learning, we demonstrate that such an approach has the potential to achieve strong performance in the future and is worthwhile for further research.