TY - GEN
T1 - Interpretable adversarial perturbation in input embedding space for text
AU - Sato, Motoki
AU - Suzuki, Jun
AU - Shindo, Hiroyuki
AU - Matsumoto, Yuji
N1 - Publisher Copyright:
© 2018 International Joint Conferences on Artificial Intelligence.All right reserved.
Copyright:
Copyright 2020 Elsevier B.V., All rights reserved.
PY - 2018
Y1 - 2018
N2 - Following great success in the image processing field, the idea of adversarial training has been applied to tasks in the natural language processing (NLP) field. One promising approach directly applies adversarial training developed in the image processing field to the input word embedding space instead of the discrete input space of texts. However, this approach abandons such interpretability as generating adversarial texts to significantly improve the performance of NLP tasks. This paper restores interpretability to such methods by restricting the directions of perturbations toward the existing words in the input embedding space. As a result, we can straightforwardly reconstruct each input with perturbations to an actual text by considering the perturbations to be the replacement of words in the sentence while maintaining or even improving the task performance.
AB - Following great success in the image processing field, the idea of adversarial training has been applied to tasks in the natural language processing (NLP) field. One promising approach directly applies adversarial training developed in the image processing field to the input word embedding space instead of the discrete input space of texts. However, this approach abandons such interpretability as generating adversarial texts to significantly improve the performance of NLP tasks. This paper restores interpretability to such methods by restricting the directions of perturbations toward the existing words in the input embedding space. As a result, we can straightforwardly reconstruct each input with perturbations to an actual text by considering the perturbations to be the replacement of words in the sentence while maintaining or even improving the task performance.
UR - http://www.scopus.com/inward/record.url?scp=85055673732&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85055673732&partnerID=8YFLogxK
U2 - 10.24963/ijcai.2018/601
DO - 10.24963/ijcai.2018/601
M3 - Conference contribution
AN - SCOPUS:85055673732
T3 - IJCAI International Joint Conference on Artificial Intelligence
SP - 4323
EP - 4330
BT - Proceedings of the 27th International Joint Conference on Artificial Intelligence, IJCAI 2018
A2 - Lang, Jerome
PB - International Joint Conferences on Artificial Intelligence
T2 - 27th International Joint Conference on Artificial Intelligence, IJCAI 2018
Y2 - 13 July 2018 through 19 July 2018
ER -