Enregistré dans:
Détails bibliographiques
Auteurs principaux: Lu, Xiaolei, Ma, Jianghong
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2404.00140
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866913292454199296
author Lu, Xiaolei
Ma, Jianghong
author_facet Lu, Xiaolei
Ma, Jianghong
contents Explainability algorithms aimed at interpreting decision-making AI systems usually consider balancing two critical dimensions: 1) \textit{faithfulness}, where explanations accurately reflect the model's inference process. 2) \textit{plausibility}, where explanations are consistent with domain experts. However, the question arises: do faithfulness and plausibility inherently conflict? In this study, through a comprehensive quantitative comparison between the explanations from the selected explainability methods and expert-level interpretations across three NLP tasks: sentiment analysis, intent detection, and topic labeling, we demonstrate that traditional perturbation-based methods Shapley value and LIME could attain greater faithfulness and plausibility. Our findings suggest that rather than optimizing for one dimension at the expense of the other, we could seek to optimize explainability algorithms with dual objectives to achieve high levels of accuracy and user accessibility in their explanations.
format Preprint
id arxiv_https___arxiv_org_abs_2404_00140
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Does Faithfulness Conflict with Plausibility? An Empirical Study in Explainable AI across NLP Tasks
Lu, Xiaolei
Ma, Jianghong
Artificial Intelligence
Machine Learning
Explainability algorithms aimed at interpreting decision-making AI systems usually consider balancing two critical dimensions: 1) \textit{faithfulness}, where explanations accurately reflect the model's inference process. 2) \textit{plausibility}, where explanations are consistent with domain experts. However, the question arises: do faithfulness and plausibility inherently conflict? In this study, through a comprehensive quantitative comparison between the explanations from the selected explainability methods and expert-level interpretations across three NLP tasks: sentiment analysis, intent detection, and topic labeling, we demonstrate that traditional perturbation-based methods Shapley value and LIME could attain greater faithfulness and plausibility. Our findings suggest that rather than optimizing for one dimension at the expense of the other, we could seek to optimize explainability algorithms with dual objectives to achieve high levels of accuracy and user accessibility in their explanations.
title Does Faithfulness Conflict with Plausibility? An Empirical Study in Explainable AI across NLP Tasks
topic Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2404.00140