English

Efficient Decompositional Rule Extraction for Deep Neural Networks

Machine Learning 2021-11-25 v1 Artificial Intelligence

Abstract

In recent years, there has been significant work on increasing both interpretability and debuggability of a Deep Neural Network (DNN) by extracting a rule-based model that approximates its decision boundary. Nevertheless, current DNN rule extraction methods that consider a DNN's latent space when extracting rules, known as decompositional algorithms, are either restricted to single-layer DNNs or intractable as the size of the DNN or data grows. In this paper, we address these limitations by introducing ECLAIRE, a novel polynomial-time rule extraction algorithm capable of scaling to both large DNN architectures and large training datasets. We evaluate ECLAIRE on a wide variety of tasks, ranging from breast cancer prognosis to particle detection, and show that it consistently extracts more accurate and comprehensible rule sets than the current state-of-the-art methods while using orders of magnitude less computational resources. We make all of our methods available, including a rule set visualisation interface, through the open-source REMIX library (https://github.com/mateoespinosa/remix).

Keywords

Cite

@article{arxiv.2111.12628,
  title  = {Efficient Decompositional Rule Extraction for Deep Neural Networks},
  author = {Mateo Espinosa Zarlenga and Zohreh Shams and Mateja Jamnik},
  journal= {arXiv preprint arXiv:2111.12628},
  year   = {2021}
}

Comments

Accepted at NeurIPS 2021 Workshop on eXplainable AI approaches for debugging and diagnosis (XAI4Debugging)

R2 v1 2026-06-24T07:50:52.122Z