The Quarrel of Local Post-hoc Explainers for Moral Values Classification in Natural Language Processing

The Quarrel of Local Post-hoc Explainers for Moral Values Classification in Natural Language Processing

Agiollo, A. (TU Delft Interactive Intelligence; Alma Mater Studiorum – Universitá di Bologna)
Cavalcante Siebert, L. (TU Delft Interactive Intelligence)
Murukannaiah, P.K. (TU Delft Interactive Intelligence)
Omicini, Andrea (Alma Mater Studiorum – Universitá di Bologna)

Calvaresi, Davide (editor)
Najjar, Amro (editor)
Omicini, Andrea (editor)
Carli, Rachele (editor)
Ciatto, Giovanni (editor)
Aydogan, Reyhan (editor)
Mualla, Yazan (editor)
Främling, Kary (editor)

2023

Although popular and effective, large language models (LLM) are characterised by a performance vs. transparency trade-off that hinders their applicability to sensitive scenarios. This is the main reason behind many approaches focusing on local post-hoc explanations recently proposed by the XAI community. However, to the best of our knowledge, a thorough comparison among available explainability techniques is currently missing, mainly for the lack of a general metric to measure their benefits. We compare state-of-the-art local post-hoc explanation mechanisms for models trained over moral value classification tasks based on a measure of correlation. By relying on a novel framework for comparing global impact scores, our experiments show how most local post-hoc explainers are loosely correlated, and highlight huge discrepancies in their results—their “quarrel” about explanations. Finally, we compare the impact scores distribution obtained from each local post-hoc explainer with human-made dictionaries, and point out that there is no correlation between explanation outputs and the concepts humans consider as salient.

eXplainable Artificial Intelligence
Local Post-hoc Explanations
Moral Values Classification
Natural Language Processing

http://resolver.tudelft.nl/uuid:bcccfa5e-f406-4ac6-8cf2-71eb00b07d07

https://doi.org/10.1007/978-3-031-40878-6_6

Springer

2024-04-01

9783031408779

Explainable and Transparent AI and Multi-Agent Systems - 5th International Workshop, EXTRAAMAS 2023, Revised Selected Papers

Proceedings of the 5th International Workshop on EXTRAAMAS 2023, 2023-05-29, London, United Kingdom

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 0302-9743, 14127 LNAI

Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Institutional Repository

conference paper

© 2023 A. Agiollo, L. Cavalcante Siebert, P.K. Murukannaiah, Andrea Omicini