Simplified Planner Selection

Anonymous

Keywords: Planning, Machine Learning, Portfolio

TL;DR: Investigating into the Delfi approach for planner selection and showing that simpler and understandable models can outperform previous aproaches.

Abstract:

There exists no planning algorithm that outperforms all others.
Therefore, it is important to know which algorithm works
well on a task. A recently published approach uses either image
or graph convolutional neural networks to solve this problem
and achieves top performance. Especially the transformation
from the task to an image ignores a lot of information.
Thus, we would like to know what the network is learning
and if this is reasonable. As this is currently not possible, we
take one step back. We identify a small set of simple graph
features and show that elementary and interpretable machine
learning techniques can use those features to outperform the
neural network based approach. Furthermore, we evaluate the
importance of those features and verify that the performance
of our approach is robust to changes in the training and test
data.