A Comparative Analysis of Counterfactual Explanation Methods for Text Classifiers
This work addresses the need for interpretable text classifiers, but it is incremental as it evaluates existing methods without introducing new ones.
The paper compared five counterfactual explanation methods for text classifiers, finding that established white-box methods effectively change classifier outputs, while newer LLM-based methods produce more natural text but often fail to alter outputs.
Counterfactual explanations can be used to interpret and debug text classifiers by producing minimally altered text inputs that change a classifier's output. In this work, we evaluate five methods for generating counterfactual explanations for a BERT text classifier on two datasets using three evaluation metrics. The results of our experiments suggest that established white-box substitution-based methods are effective at generating valid counterfactuals that change the classifier's output. In contrast, newer methods based on large language models (LLMs) excel at producing natural and linguistically plausible text counterfactuals but often fail to generate valid counterfactuals that alter the classifier's output. Based on these results, we recommend developing new counterfactual explanation methods that combine the strengths of established gradient-based approaches and newer LLM-based techniques to generate high-quality, valid, and plausible text counterfactual explanations.