English

EVO-LRP: Evolutionary Optimization of LRP for Interpretable Model Explanations

Machine Learning 2025-10-02 v2 Artificial Intelligence Computer Vision and Pattern Recognition

Abstract

Explainable AI (XAI) methods help identify which image regions influence a model's prediction, but often face a trade-off between detail and interpretability. Layer-wise Relevance Propagation (LRP) offers a model-aware alternative. However, LRP implementations commonly rely on heuristic rule sets that are not optimized for clarity or alignment with model behavior. We introduce EVO-LRP, a method that applies Covariance Matrix Adaptation Evolution Strategy (CMA-ES) to tune LRP hyperparameters based on quantitative interpretability metrics, such as faithfulness or sparseness. EVO-LRP outperforms traditional XAI approaches in both interpretability metric performance and visual coherence, with strong sensitivity to class-specific features. These findings demonstrate that attribution quality can be systematically improved through principled, task-specific optimization.

Keywords

Cite

@article{arxiv.2509.23585,
  title  = {EVO-LRP: Evolutionary Optimization of LRP for Interpretable Model Explanations},
  author = {Emerald Zhang and Julian Weaver and Samantha R Santacruz and Edward Castillo},
  journal= {arXiv preprint arXiv:2509.23585},
  year   = {2025}
}

Comments

15 pages

R2 v1 2026-07-01T06:01:50.373Z