Improving Behavioural Cloning with Human-Driven Dynamic Dataset Augmentation
This addresses limitations in behavioral cloning for training agents, though it appears incremental as it builds on existing methods with human feedback.
The paper tackles the problem of behavioral cloning's dependence on expert data distribution by introducing human-in-the-loop corrections during simulation, resulting in better policies with improved quantitative evaluation and human-likeliness.
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.