Adversarial Edit Attacks for Tree Data

TL;DR

This paper introduces a novel black-box adversarial attack method for tree-structured data, demonstrating its effectiveness on machine learning models in medicine and program analysis without requiring gradient information.

Contribution

It extends adversarial attack techniques to tree data using tree edit distance and black-box queries, a novel approach in this domain.

Findings

Effective attacks on tree classifiers like tree-kernel-SVMs and recursive neural networks.

Requires only logarithmic black-box queries and no gradient information.

Applicable to biomedical and programming data sets.

Abstract

Many machine learning models can be attacked with adversarial examples, i.e. inputs close to correctly classified examples that are classified incorrectly. However, most research on adversarial attacks to date is limited to vectorial data, in particular image data. In this contribution, we extend the field by introducing adversarial edit attacks for tree-structured data with potential applications in medicine and automated program analysis. Our approach solely relies on the tree edit distance and a logarithmic number of black-box queries to the attacked classifier without any need for gradient information. We evaluate our approach on two programming and two biomedical data sets and show that many established tree classifiers, like tree-kernel-SVMs and recursive neural networks, can be attacked effectively.