Explaining Classifiers Trained on Raw Hierarchical Multiple-Instance Data

TL;DR

This paper introduces a method for explaining classifiers trained on raw structured hierarchical data, improving interpretability and computational efficiency over existing techniques.

Contribution

It presents a novel approach to generate interpretable explanations for classifiers on raw hierarchical data, with faster and higher-quality results than prior methods.

Findings

Order of magnitude faster explanation generation

Higher-quality explanations compared to graph neural network methods

Effective interpretation of classifiers trained on raw structured data

Abstract

Learning from raw data input, thus limiting the need for feature engineering, is a component of many successful applications of machine learning methods in various domains. While many problems naturally translate into a vector representation directly usable in standard classifiers, a number of data sources have the natural form of structured data interchange formats (e.g., security logs in JSON/XML format). Existing methods, such as in Hierarchical Multiple Instance Learning (HMIL), allow learning from such data in their raw form. However, the explanation of the classifiers trained on raw structured data remains largely unexplored. By treating these models as sub-set selections problems, we demonstrate how interpretable explanations, with favourable properties, can be generated using computationally efficient algorithms. We compare to an explanation technique adopted from graph neural networks showing an order of magnitude speed-up and higher-quality explanations.