Are Your Explanations Reliable? Investigating the Stability of LIME in Explaining Text Classifiers by Marrying XAI and Adversarial Attack
This addresses reliability issues in explainable AI for critical applications like healthcare and finance, but it is incremental as it builds on existing XAI and adversarial attack methods.
The paper tackles the problem of LIME's instability in explaining text classifiers by proposing XAIFooler, a novel algorithm that perturbs text inputs to manipulate explanations while preserving semantics and predictions, with experiments showing it significantly outperforms baselines in manipulation ability.
LIME has emerged as one of the most commonly referenced tools in explainable AI (XAI) frameworks that is integrated into critical machine learning applications--e.g., healthcare and finance. However, its stability remains little explored, especially in the context of text data, due to the unique text-space constraints. To address these challenges, in this paper, we first evaluate the inherent instability of LIME on text data to establish a baseline, and then propose a novel algorithm XAIFooler to perturb text inputs and manipulate explanations that casts investigation on the stability of LIME as a text perturbation optimization problem. XAIFooler conforms to the constraints to preserve text semantics and original prediction with small perturbations, and introduces Rank-biased Overlap (RBO) as a key part to guide the optimization of XAIFooler that satisfies all the requirements for explanation similarity measure. Extensive experiments on real-world text datasets demonstrate that XAIFooler significantly outperforms all baselines by large margins in its ability to manipulate LIME's explanations with high semantic preservability.