Counterfactuals of Counterfactuals: a back-translation-inspired approach to analyse counterfactual editors
This work addresses the challenge of evaluating interpretability methods for NLP models, which is crucial for responsible AI, though it is incremental as it builds on existing counterfactual explanation techniques.
The paper tackles the problem of evaluating counterfactual explanation methods in NLP by proposing a back-translation-inspired approach that uses earlier outputs as ground truth proxies to analyze consistency, resulting in a novel metric for assessing explainer behavior.
In the wake of responsible AI, interpretability methods, which attempt to provide an explanation for the predictions of neural models have seen rapid progress. In this work, we are concerned with explanations that are applicable to natural language processing (NLP) models and tasks, and we focus specifically on the analysis of counterfactual, contrastive explanations. We note that while there have been several explainers proposed to produce counterfactual explanations, their behaviour can vary significantly and the lack of a universal ground truth for the counterfactual edits imposes an insuperable barrier on their evaluation. We propose a new back translation-inspired evaluation methodology that utilises earlier outputs of the explainer as ground truth proxies to investigate the consistency of explainers. We show that by iteratively feeding the counterfactual to the explainer we can obtain valuable insights into the behaviour of both the predictor and the explainer models, and infer patterns that would be otherwise obscured. Using this methodology, we conduct a thorough analysis and propose a novel metric to evaluate the consistency of counterfactual generation approaches with different characteristics across available performance indicators.