Maximum Entropy Deep Inverse Reinforcement Learning

TL;DR

This paper introduces a deep neural network framework for inverse reinforcement learning based on maximum entropy principles, enabling efficient training and application to complex, nonlinear reward functions with improved performance on benchmarks.

Contribution

It develops a novel deep IRL approach leveraging maximum entropy, allowing for scalable, efficient learning of complex reward functions with minimal dependence on precomputed features.

Findings

Achieves performance comparable to state-of-the-art on standard benchmarks.

Outperforms on a benchmark with highly varying reward structures.

Extends architecture to include larger convolutions for raw input processing.

Abstract

This paper presents a general framework for exploiting the representational capacity of neural networks to approximate complex, nonlinear reward functions in the context of solving the inverse reinforcement learning (IRL) problem. We show in this context that the Maximum Entropy paradigm for IRL lends itself naturally to the efficient training of deep architectures. At test time, the approach leads to a computational complexity independent of the number of demonstrations, which makes it especially well-suited for applications in life-long learning scenarios. Our approach achieves performance commensurate to the state-of-the-art on existing benchmarks while exceeding on an alternative benchmark based on highly varying reward structures. Finally, we extend the basic architecture - which is equivalent to a simplified subclass of Fully Convolutional Neural Networks (FCNNs) with width one - to include larger convolutions in order to eliminate dependency on precomputed spatial features and work on raw input representations.