TY - GEN
T1 - Evaluating and Enhancing the Robustness of Retrieval-Based Dialogue Systems with Adversarial Examples
AU - Li, Jia
AU - Tao, Chongyang
AU - Peng, Nanyun
AU - Wu, Wei
AU - Zhao, Dongyan
AU - Yan, Rui
N1 - Publisher Copyright:
© 2019, Springer Nature Switzerland AG.
PY - 2019
Y1 - 2019
N2 - Retrieval-based dialogue systems have shown strong performances on both consistency and fluency according to several recent studies. However, their robustness towards malicious attacks remains largely untested. In this paper, we generate adversarial examples in black-box settings to evaluate the robustness of retrieval-based dialogue systems. On three representative retrieval-based dialogue models, our attacks reduce R by 38.3 45.0 and 31.5 respectively on the Ubuntu dataset. Moreover, with adversarial training using our generated adversarial examples, we significantly improve the robustness of retrieval-based dialogue systems. We conduct thorough analysis to understand the robustness of retrieval-based dialog systems. Our results provide new insights to facilitate future work on building more robust dialogue systems.
AB - Retrieval-based dialogue systems have shown strong performances on both consistency and fluency according to several recent studies. However, their robustness towards malicious attacks remains largely untested. In this paper, we generate adversarial examples in black-box settings to evaluate the robustness of retrieval-based dialogue systems. On three representative retrieval-based dialogue models, our attacks reduce R by 38.3 45.0 and 31.5 respectively on the Ubuntu dataset. Moreover, with adversarial training using our generated adversarial examples, we significantly improve the robustness of retrieval-based dialogue systems. We conduct thorough analysis to understand the robustness of retrieval-based dialog systems. Our results provide new insights to facilitate future work on building more robust dialogue systems.
KW - Adversarial examples
KW - Retrieval-based dialogue systems
UR - https://www.scopus.com/pages/publications/85075554699
U2 - 10.1007/978-3-030-32233-5_12
DO - 10.1007/978-3-030-32233-5_12
M3 - 会议稿件
AN - SCOPUS:85075554699
SN - 9783030322328
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 142
EP - 154
BT - Natural Language Processing and Chinese Computing - 8th CCF International Conference, NLPCC 2019, Proceedings
A2 - Tang, Jie
A2 - Kan, Min-Yen
A2 - Zhao, Dongyan
A2 - Li, Sujian
A2 - Zan, Hongying
PB - Springer
T2 - 8th CCF International Conference on Natural Language Processing and Chinese Computing, NLPCC 2019
Y2 - 9 October 2019 through 14 October 2019
ER -