Argot: Generating Adversarial Readable Chinese Texts

Natural language processing (NLP) models are known vulnerable to adversarial examples, similar to image processing models. Studying adversarial texts is an essential step to improve the robustness of NLP models. However, existing studies mainly focus on analyzing English texts and generating adversarial examples for English texts. There is no work studying the possibility and effect of the transformation to another language, e.g, Chinese. In this paper, we analyze the differences between Chinese and English, and explore the methodology to transform the existing English adversarial generation method to Chinese. We propose a novel black-box adversarial Chinese texts generation solution Argot, by utilizing the method for adversarial English samples and several novel methods developed on Chinese characteristics. Argot could effectively and efficiently generate adversarial Chinese texts with good readability. Furthermore, Argot could also automatically generate targeted Chinese adversarial text, achieving a high success rate and ensuring readability of the Chinese.

[1]  Mani B. Srivastava,et al.  Generating Natural Language Adversarial Examples , 2018, EMNLP.

[2]  Jeffrey Dean,et al.  Efficient Estimation of Word Representations in Vector Space , 2013, ICLR.

[3]  Michael I. Jordan,et al.  Advances in Neural Information Processing Systems 30 , 1995 .

[4]  Xiang Zhang,et al.  Character-level Convolutional Networks for Text Classification , 2015, NIPS.

[5]  P. Cochat,et al.  Et al , 2008, Archives de pediatrie : organe officiel de la Societe francaise de pediatrie.

[6]  Vadlamani Ravi,et al.  A survey on opinion mining and sentiment analysis: Tasks, approaches and applications , 2015, Knowl. Based Syst..

[7]  Yoon Kim,et al.  Convolutional Neural Networks for Sentence Classification , 2014, EMNLP.

[8]  Andrew M. Dai,et al.  Adversarial Training Methods for Semi-Supervised Text Classification , 2016, ICLR.

[9]  Eric Gilbert,et al.  Algorithmically Bypassing Censorship on Sina Weibo with Nondeterministic Homophone Substitutions , 2015, ICWSM.

[10]  Hitoshi Iida.  Proceedings of the 38th Annual Meeting on Association for Computational Linguistics , 2000 .

[11]  Dejing Dou,et al.  HotFlip: White-Box Adversarial Examples for Text Classification , 2017, ACL.

[12]  Bo Li,et al.  Adversarial Texts with Gradient Methods , 2018, ArXiv.

[13]  Wei Wu,et al.  Glyce: Glyph-vectors for Chinese Character Representations , 2019, NeurIPS.

[14]  Yanjun Qi,et al.  Black-Box Generation of Adversarial Text Sequences to Evade Deep Learning Classifiers , 2018, 2018 IEEE Security and Privacy Workshops (SPW).

[15]  Sameep Mehta,et al.  Towards Crafting Text Adversarial Samples , 2017, ArXiv.

[16]  Joel R. Tetreault,et al.  Abusive Language Detection in Online User Content , 2016, WWW.

[17]  Yann LeCun,et al.  Learning a similarity metric discriminatively, with application to face verification , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[18]  Ting Wang,et al.  TextBugger: Generating Adversarial Text Against Real-world Applications , 2018, NDSS.

[19]  Xiaoyong Du,et al.  Analogical Reasoning on Chinese Morphological and Semantic Relations , 2018, ACL.

[20]  Hiroyuki Shindo,et al.  Interpretable Adversarial Perturbation in Input Embedding Space for Text , 2018, IJCAI.

[21]  Xin Jin,et al.  Improve word embedding using both writing and pronunciation , 2018, PloS one.