Generating and Modifying Natural Language Explanations

Abdus Salam, Rolf Schwitter, Mehmet Orgun


Abstract
HESIP is a hybrid explanation system for image predictions that combines sub-symbolic and symbolic machine learning techniques to explain the predictions of image classification tasks. The sub-symbolic component makes a prediction for an image and the symbolic component learns probabilistic symbolic rules in order to explain that prediction. In HESIP, the explanations are generated in controlled natural language from the learned probabilistic rules using a bi-directional logic grammar. In this paper, we present an explanation modification method where a human-in-the-loop can modify an incorrect explanation generated by the HESIP system and afterwards, the modified explanation is used by HESIP to learn a better explanation.
Anthology ID:
2021.alta-1.15
Volume:
Proceedings of the The 19th Annual Workshop of the Australasian Language Technology Association
Month:
December
Year:
2021
Address:
Online
Venue:
ALTA
SIG:
Publisher:
Australasian Language Technology Association
Note:
Pages:
149–157
Language:
URL:
https://aclanthology.org/2021.alta-1.15
DOI:
Bibkey:
Cite (ACL):
Abdus Salam, Rolf Schwitter, and Mehmet Orgun. 2021. Generating and Modifying Natural Language Explanations. In Proceedings of the The 19th Annual Workshop of the Australasian Language Technology Association, pages 149–157, Online. Australasian Language Technology Association.
Cite (Informal):
Generating and Modifying Natural Language Explanations (Salam et al., ALTA 2021)
Copy Citation:
PDF:
https://preview.aclanthology.org/update-css-js/2021.alta-1.15.pdf