Neurosymbolic Imitation Learning with Human Guidance: A Privileged Information Approach

TL;DR

This paper introduces a neurosymbolic imitation learning method that leverages privileged information like gaze data during training to improve generalization and efficiency in complex environments.

Contribution

It presents a novel neurosymbolic approach that combines neural and symbolic methods, effectively utilizing privileged training data for better imitation learning.

Findings

Demonstrates improved generalization over purely neural methods

Shows efficiency gains in training with privileged information

Validates approach through empirical evaluations

Abstract

Imitation learning is widely used for learning to act in complex environments. While pure neural-based methods handle high dimensional data effectively, they suffer from the requirement of large number of samples and are prone to overfitting. Pure symbolic approaches, while generalize well, do not handle high-dimensional data effectively. We propose a neurosymbolic approach that achieves the best of both worlds, i.e, handling high-dimensional data while achieving generalization. The key advantage of our approach is that it can effectively exploit additional privileged information that is available only during training (in our case, gaze data). Our empirical evaluations demonstrate the effectiveness, efficiency and the generalization capability of our proposed approach.