Abstract
Adversarial purification is a defense mechanism for safeguarding classifiers against adversarial attacks without knowing the type of attacks or training of the classifier. These techniques characterize and eliminate adversarial perturbations from the attacked inputs, aiming to restore purified samples that retain similarity to the initially attacked ones and are correctly classified by the classifier. Due to the inherent challenges associated with characterizing noise perturbations for discrete inputs, adversarial text purification has been relatively unexplored. In this paper, we investigate the effectiveness of adversarial purification methods in defending text classifiers. We propose a novel adversarial text purification that harnesses the generative capabilities of Large Language Models (LLMs) to purify adversarial text without the need to explicitly characterize the discrete noise perturbations. We utilize prompt engineering to exploit LLMs for recovering the purified samples for given adversarial examples such that they are semantically similar and correctly classified. Our proposed method demonstrates remarkable performance over various classifiers, improving their accuracy under the attack by over 65% on average.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Alizadeh, M., et al.: Open-source large language models outperform crowd workers and approach ChatGPT in text-annotation tasks. arXiv preprint arXiv:2307.02179 (2023)
Alzantot, M., Sharma, Y., Elgohary, A., Ho, B.J., Srivastava, M., Chang, K.W.: Generating natural language adversarial examples. arXiv preprint arXiv:1804.07998 (2018)
Bai, Y., , et al.: Training a helpful and harmless assistant with reinforcement learning from human feedback. arXiv preprint arXiv:2204.05862 (2022)
Bhattacharjee, A., Liu, H.: Fighting fire with fire: can ChatGPT detect AI-generated text? arXiv preprint arXiv:2308.01284 (2023)
Bhattacharjee, A., Moraffah, R., Garland, J., Liu, H.: LLMS as counterfactual explanation modules: can ChatGPT explain black-box text classifiers? arXiv preprint arXiv:2309.13340 (2023)
Brown, T., et al.: Language models are few-shot learners. Adv. Neural. Inf. Process. Syst. 33, 1877–1901 (2020)
Chang, Y., et al.: A survey on evaluation of large language models. arXiv preprint arXiv:2307.03109 (2023)
Cheng, Y., Jiang, L., Macherey, W.: Robust neural machine translation with doubly adversarial inputs. arXiv preprint arXiv:1906.02443 (2019)
Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: BERT: pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805 (2018)
Ebrahimi, J., Rao, A., Lowd, D., Dou, D.: HotFlip: white-box adversarial examples for text classification. arXiv preprint arXiv:1712.06751 (2017)
Flamholz, Z.N., Biller, S.J., Kelly, L.: Large language models improve annotation of viral proteins. Res. Sq. (2023)
Jia, R., Raghunathan, A., Göksel, K., Liang, P.: Certified robustness to adversarial word substitutions. arXiv preprint arXiv:1909.00986 (2019)
Jin, D., Jin, Z., Zhou, J.T., Szolovits, P.: Is BERT really robust? A strong baseline for natural language attack on text classification and entailment. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 34, pp. 8018–8025 (2020)
Latif, S., Usama, M., Malik, M.I., Schuller, B.W.: Can large language models aid in annotating speech emotional data? uncovering new frontiers. arXiv preprint arXiv:2307.06090 (2023)
LeCun, Y., Chopra, S., Hadsell, R., Ranzato, M., Huang, F.: A tutorial on energy-based learning. Predicting Struct. Data 1(0) (2006)
Li, L., Ma, R., Guo, Q., Xue, X., Qiu, X.: BERT-attack: adversarial attack against BERT using BERT. arXiv preprint arXiv:2004.09984 (2020)
Li, L., Qiu, X.: Token-aware virtual adversarial training in natural language understanding. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 35, pp. 8410–8418 (2021)
Li, L., Song, D., Qiu, X.: Text adversarial purification as defense against adversarial attacks. arXiv preprint arXiv:2203.14207 (2022)
Li, Z., et al.: Searching for an effective defender: benchmarking defense against adversarial word substitution. arXiv preprint arXiv:2108.12777 (2021)
Liu, Y., et al.: RoBERTa: a robustly optimized BERT pretraining approach. arXiv preprint arXiv:1907.11692 (2019)
Maas, A., et al.: Learning word vectors for sentiment analysis. In: Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, pp. 142–150 (2011)
Madry, A., Makelov, A., Schmidt, L., Tsipras, D., Vladu, A.: Towards deep learning models resistant to adversarial attacks. arXiv preprint arXiv:1706.06083 (2017)
Miyato, T., Dai, A.M., Goodfellow, I.: Adversarial training methods for semi-supervised text classification. arXiv preprint arXiv:1605.07725 (2016)
Morris, J.X., Lifland, E., Yoo, J.Y., Grigsby, J., Jin, D., Qi, Y.: TextAttack: a framework for adversarial attacks, data augmentation, and adversarial training in NLP. arXiv preprint arXiv:2005.05909 (2020)
Nie, W., Guo, B., Huang, Y., Xiao, C., Vahdat, A., Anandkumar, A.: Diffusion models for adversarial purification. arXiv preprint arXiv:2205.07460 (2022)
Ouyang, L., et al.: Training language models to follow instructions with human feedback. Adv. Neural. Inf. Process. Syst. 35, 27730–27744 (2022)
Peng, C., et al.: A study of generative large language model for medical research and healthcare. arXiv preprint arXiv:2305.13523 (2023)
Radford, A., et al.: Language models are unsupervised multitask learners. OpenAI Blog 1(8), 9 (2019)
Ren, S., Deng, Y., He, K., Che, W.: Generating natural language adversarial examples through probability weighted word saliency. In: Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pp. 1085–1097 (2019)
Samangouei, P., Kabkab, M., Chellappa, R.: Defense-GAN: protecting classifiers against adversarial attacks using generative models. arXiv preprint arXiv:1805.06605 (2018)
Shi, C., Holtz, C., Mishne, G.: Online adversarial purification based on self-supervision. arXiv preprint arXiv:2101.09387 (2021)
Shreya, G., Khapra, M.M.: A survey in adversarial defences and robustness in NLP. arXiv preprint arXiv:2203.06414 (2022)
Song, Y., Sohl-Dickstein, J., Kingma, D.P., Kumar, A., Ermon, S., Poole, B.: Score-based generative modeling through stochastic differential equations. arXiv preprint arXiv:2011.13456 (2020)
Xu, B., et al.: Expertprompting: instructing large language models to be distinguished experts. arXiv preprint arXiv:2305.14688 (2023)
Ye, M., Miao, C., Wang, T., Ma, F.: TextHoaxer: budgeted hard-label adversarial attacks on text. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 36, pp. 3877–3884 (2022)
Yoon, J., Hwang, S.J., Lee, J.: Adversarial purification with score-based generative models. In: International Conference on Machine Learning. pp. 12062–12072. PMLR (2021)
Zeng, J., Xu, J., Zheng, X., Huang, X.: Certified robustness to text adversarial attacks by randomized [mask]. Comput. Linguist. 49(2), 395–427 (2023)
Zhu, C., Cheng, Y., Gan, Z., Sun, S., Goldstein, T., Liu, J.: Freelb: Enhanced adversarial training for natural language understanding. arXiv preprint arXiv:1909.11764 (2019)
Acknowledgements
This work is supported by Army Research Office (ARO) W911NF2110030 and Army Research Laboratory (ARL) W911NF2020124. Opinions, interpretations, conclusions, and recommendations are those of the authors’ and should not be interpreted as representing the official views or policies of the Army Research Office or the Army Research Lab.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2024 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Moraffah, R., Khandelwal, S., Bhattacharjee, A., Liu, H. (2024). Adversarial Text Purification: A Large Language Model Approach for Defense. In: Yang, DN., Xie, X., Tseng, V.S., Pei, J., Huang, JW., Lin, J.CW. (eds) Advances in Knowledge Discovery and Data Mining. PAKDD 2024. Lecture Notes in Computer Science(), vol 14649. Springer, Singapore. https://doi.org/10.1007/978-981-97-2262-4_6
Download citation
DOI: https://doi.org/10.1007/978-981-97-2262-4_6
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-97-2264-8
Online ISBN: 978-981-97-2262-4
eBook Packages: Computer ScienceComputer Science (R0)