Improving Behavioural Cloning with Human-Driven Dynamic Dataset Augmentation

TL;DR

This paper introduces a human-in-the-loop method to enhance behavioural cloning by allowing experts to intervene during training, resulting in improved policies and more human-like agent behavior.

Contribution

A novel approach enabling experts to dynamically intervene during behavioural cloning, improving policy quality and training efficiency.

Findings

Enhanced policy performance in quantitative metrics

Increased human-likeliness of agent behavior

Faster training with resource-efficient corrections

Abstract

Behavioural cloning has been extensively used to train agents and is recognized as a fast and solid approach to teach general behaviours based on expert trajectories. Such method follows the supervised learning paradigm and it strongly depends on the distribution of the data. In our paper, we show how combining behavioural cloning with human-in-the-loop training solves some of its flaws and provides an agent task-specific corrections to overcome tricky situations while speeding up the training time and lowering the required resources. To do this, we introduce a novel approach that allows an expert to take control of the agent at any moment during a simulation and provide optimal solutions to its problematic situations. Our experiments show that this approach leads to better policies both in terms of quantitative evaluation and in human-likeliness.