Double Trouble: How to not explain a text classifier's decisions using counterfactuals synthesized by masked language models?
This work addresses the problem of unreliable explanation methods for machine learning practitioners, revealing critical flaws in widely used attribution techniques, making it an incremental but important critique.
The paper challenges the effectiveness of Input Marginalization (IM) for explaining text classifier decisions, showing through rigorous evaluation with 6 metrics and 3 datasets that IM is not better than a Leave-One-Out baseline, and identifies biases in common deletion-based metrics.
A principle behind dozens of attribution methods is to take the prediction difference between before-and-after an input feature (here, a token) is removed as its attribution. A popular Input Marginalization (IM) method (Kim et al., 2020) uses BERT to replace a token, yielding more plausible counterfactuals. While Kim et al. (2020) reported that IM is effective, we find this conclusion not convincing as the DeletionBERT metric used in their paper is biased towards IM. Importantly, this bias exists in Deletion-based metrics, including Insertion, Sufficiency, and Comprehensiveness. Furthermore, our rigorous evaluation using 6 metrics and 3 datasets finds no evidence that IM is better than a Leave-One-Out (LOO) baseline. We find two reasons why IM is not better than LOO: (1) deleting a single word from the input only marginally reduces a classifier's accuracy; and (2) a highly predictable word is always given near-zero attribution, regardless of its true importance to the classifier. In contrast, making LIME samples more natural via BERT consistently improves LIME accuracy under several ROAR metrics.