Verifiable Learning for Robust Tree Ensembles

TL;DR

This paper introduces large-spread decision tree ensembles and a training method for them, enabling efficient verification of robustness against evasion attacks, and demonstrating improved security and comparable accuracy.

Contribution

The paper proposes a new class of decision tree ensembles called large-spread ensembles and a training algorithm to learn them, facilitating polynomial-time verifiability of robustness.

Findings

Large-spread ensembles can be verified in seconds on standard hardware.

They are more robust against evasion attacks than traditional ensembles.

Training these ensembles incurs only a slight accuracy loss in non-adversarial settings.

Abstract

Verifying the robustness of machine learning models against evasion attacks at test time is an important research problem. Unfortunately, prior work established that this problem is NP-hard for decision tree ensembles, hence bound to be intractable for specific inputs. In this paper, we identify a restricted class of decision tree ensembles, called large-spread ensembles, which admit a security verification algorithm running in polynomial time. We then propose a new approach called verifiable learning, which advocates the training of such restricted model classes which are amenable for efficient verification. We show the benefits of this idea by designing a new training algorithm that automatically learns a large-spread decision tree ensemble from labelled data, thus enabling its security verification in polynomial time. Experimental results on public datasets confirm that large-spread ensembles trained using our algorithm can be verified in a matter of seconds, using standard commercial hardware. Moreover, large-spread ensembles are more robust than traditional ensembles against evasion attacks, at the cost of an acceptable loss of accuracy in the non-adversarial setting.