Fetching the paper…

Reading the bibliography…

2021

Defending Against Backdoor Attacks in Natural Language Generation

Sun, Xiaofei, Li, Xiaoya, Meng, Yuxian et al.

Original
PDF
Post on X

Understand

The frustratingly fragile nature of neural network models make current natural language generation (NLG) systems prone to backdoor attacks and generate malicious sequences that could be sexist or offensive.

Unfortunately, little effort has been invested to how backdoor attacks can affect current NLG models and how to defend against these attacks.
In this work, by giving a formal definition of backdoor attack and defense, we investigate this problem on two important NLG tasks, machine translation and dialog generation.
Tailored to the inherent nature of NLG models (e.g., producing a sequence of coherent words given contexts), we design defending strategies against attacks.

Defending Against Backdoor Attacks in Natural Language Generation · Around

Built on

Language Models are Few-Shot Learners
Brown, T.; Mann, B.; Ryder, N.; Subbiah, M.; Kaplan, J. D.; Dhariwal, P.; Neelakantan, A.; Shyam, P.; Sastry, G.; Askell, A.; Agarwal, S.; Herbert-Voss, A.; Krueger, G.; Henighan, T.; Child, R.; Ramesh, A.; Ziegler, D.; Wu, J.; Winter, C.; Hesse, C.; Chen, M.; Sigler, E.; Litwin, M.; Gray, S.; Chess, B.; Clark, J.; Berner, C.; McCandlish, S.; Radford, A.; Sutskever, I.; and Amodei, D. 2020 · 1901
Earlier work this paper cites.
Bertscore: Evaluating text generation with bert
Original
Zhang, T.; Kishore, V.; Wu, F.; Weinberger, K. Q.; and Artzi, Y. 2019 · 1904
Earlier work this paper cites.
Unified language model pre-training for natural language understanding and generation
Original
Dong, L.; Yang, N.; Wang, W.; Wei, F.; Liu, X.; Wang, Y.; Gao, J.; Zhou, M.; and Hon, H.-W. 2019 · 1905
Earlier work this paper cites.
Xlnet: Generalized autoregressive pretraining for language understanding
Original
Yang, Z.; Dai, Z.; Yang, Y.; Carbonell, J.; Salakhutdinov, R.; and Le, Q. V. 2019 · 1906
Earlier work this paper cites.
Roberta: A robustly optimized bert pretraining approach
Original
Liu, Y.; Ott, M.; Goyal, N.; Du, J.; Joshi, M.; Chen, D.; Levy, O.; Lewis, M.; Zettlemoyer, L.; and Stoyanov, V. 2019 · 1907
Earlier work this paper cites.
Bart: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension
Original
Lewis, M.; Liu, Y.; Goyal, N.; Ghazvininejad, M.; Mohamed, A.; Levy, O.; Stoyanov, V.; and Zettlemoyer, L. 2019 · 1910
Earlier work this paper cites.
Defending neural backdoors via generative distribution modeling
Original
Qiao, X.; Yang, Y.; and Li, H. 2019 · 1910
Earlier work this paper cites.
Exploring the limits of transfer learning with a unified text-to-text transformer
Original
Raffel, C.; Shazeer, N.; Roberts, A.; Lee, K.; Narang, S.; Matena, M.; Zhou, Y.; Li, W.; and Liu, P. J. 2019 · 1910
Earlier work this paper cites.
Smart: Robust and efficient fine-tuning for pre-trained natural language models through principled regularized optimization
Original
Jiang, H.; He, P.; Chen, W.; Liu, X.; Gao, J.; and Zhao, T. 2019 · 1911
Earlier work this paper cites.
Speech understanding systems: A summary of results of the five-year research effort
Reddy, D. R.; et al. 1977 · 1977
Earlier work this paper cites.
Teaching Machines to Converse
Original
Li, J. 2020 · 2001
Earlier work this paper cites.
Non-autoregressive neural dialogue generation
Original
Han, Q.; Meng, Y.; Wu, F.; and Li, J. 2020 · 2002
Earlier work this paper cites.
Bleu: a method for automatic evaluation of machine translation
Papineni, K.; Roukos, S.; Ward, T.; and Zhu, W.-J. 2002 · 2002
Earlier work this paper cites.
Electra: Pre-training text encoders as discriminators rather than generators
Original
Clark, K.; Luong, M.-T.; Le, Q. V.; and Manning, C. D. 2020 · 2003
Earlier work this paper cites.
Optimus: Organizing sentences via pre-trained modeling of a latent space
Original
Li, C.; Gao, X.; Li, Y.; Peng, B.; Li, X.; Zhang, Y.; and Gao, J. 2020a · 2004
Earlier work this paper cites.
Rethinking the trigger of backdoor attack
Original
Li, Y.; Zhai, T.; Wu, B.; Jiang, Y.; Li, Z.; and Xia, S. 2020b · 2004
Earlier work this paper cites.
Rouge: A package for automatic evaluation of summaries
Lin, C.-Y. 2004 · 2004
Earlier work this paper cites.
Badnl: Backdoor attacks against nlp models
Original
Chen, X.; Salem, A.; Backes, M.; Ma, S.; and Zhang, Y. 2020 · 2006
Earlier work this paper cites.
Deberta: Decoding-enhanced bert with disentangled attention
Original
He, P.; Liu, X.; Gao, J.; and Chen, W. 2020 · 2006
Earlier work this paper cites.
Defense against Adversarial Attacks in NLP via Dirichlet Neighborhood Ensemble
Original
Zhou, Y.; Zheng, X.; Hsieh, C.-J.; Chang, K.-w.; and Huang, X. 2020 · 2006
Earlier work this paper cites.

Similar

Generating Adversarial Examples withControllable Non-transferability
Original
Wang, R.; Zhang, T.; Xie, X.; Ma, L.; Tian, C.; Juefei-Xu, F.; and Liu, Y. 2020a · 2007
Cited alongside, same era.
Big bird: Transformers for longer sequences
Original
Zaheer, M.; Guruganesh, G.; Dubey, A.; Ainslie, J.; Alberti, C.; Ontanon, S.; Pham, P.; Ravula, A.; Wang, Q.; Yang, L.; et al. 2020 · 2007
Cited alongside, same era.
DeLighT: Very Deep and Light-weight Transformer
Original
Mehta, S.; Ghazvininejad, M.; Iyer, S.; Zettlemoyer, L.; and Hajishirzi, H. 2020 · 2008
Cited alongside, same era.
Input-aware dynamic backdoor attack
Original
Nguyen, A.; and Tran, A. 2020 · 2010
Cited alongside, same era.
BAAAN: Backdoor Attacks Against Autoencoder and GAN-Based Machine Learning Models

Then

Decoding with value networks for neural machine translation
He, D.; Lu, H.; Xia, Y.; Qin, T.; Wang, L.; and Liu, T.-Y. 2017 · 2017
Later among the works it cites.
Adversarial learning for neural dialogue generation
Original
Li, J.; Monroe, W.; Shi, T.; Jean, S.; Ritter, A.; and Jurafsky, D. 2017 · 2017
Later among the works it cites.
Deep text classification can be fooled
Original
Liang, B.; Li, H.; Su, M.; Bian, P.; Li, X.; and Shi, W. 2017 · 2017
Later among the works it cites.
Generating more interesting responses in neural conversation models with distributional constraints
Original
Baheti, A.; Ritter, A.; Li, J.; and Dolan, B. 2018 · 2018
Later among the works it cites.
Bert: Pre-training of deep bidirectional transformers for language understanding
Original

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…

Salem, A.; Sautter, Y.; Backes, M.; Humbert, M.; and Zhang, Y. 2020 · 2010

Cited alongside, same era.

Onion: A simple and effective defense against textual backdoor attacks

Qi, F.; Chen, Y.; Li, M.; Liu, Z.; and Sun, M. 2020 · 2011

Cited alongside, same era.

OpenViDial: A Large-Scale, Open-Domain Dialogue Dataset with Visual Contexts

Meng, Y.; Wang, S.; Han, Q.; Sun, X.; Wu, F.; Yan, R.; and Li, J. 2020 · 2012

Cited alongside, same era.

Parallel Data, Tools and Interfaces in OPUS

Tiedemann, J. 2012 · 2012

Cited alongside, same era.

Neural machine translation by jointly learning to align and translate

Bahdanau, D.; Cho, K.; and Bengio, Y. 2014 · 2014

Cited alongside, same era.

Sequence to Sequence Learning with Neural Networks

Sutskever, I.; Vinyals, O.; and Le, Q. V. 2014 · 2014

Cited alongside, same era.

A diversity-promoting objective function for neural conversation models

Li, J.; Galley, M.; Brockett, C.; Gao, J.; and Dolan, B. 2015 · 2015

Cited alongside, same era.

Effective Approaches to Attention-based Neural Machine Translation

Luong, T.; Pham, H.; and Manning, C. D. 2015b · 2015

Cited alongside, same era.

Devlin, J.; Chang, M.-W.; Lee, K.; and Toutanova, K. 2018 · 2018

Later among the works it cites.

Neural approaches to conversational ai

Gao, J.; Galley, M.; and Li, L. 2018 · 2018

Later among the works it cites.

Interpretable Adversarial Perturbation in Input Embedding Space for Text

Sato, M.; Suzuki, J.; Shindo, H.; and Matsumoto, Y. 2018 · 2018

Later among the works it cites.

Personalizing dialogue agents: I have a dog, do you have pets too?

Zhang, S.; Dinan, E.; Urbanek, J.; Szlam, A.; Kiela, D.; and Weston, J. 2018 · 2018

Later among the works it cites.

A backdoor attack against LSTM-based text classification systems

Dai, J.; Chen, C.; and Li, Y. 2019 · 2019

Later among the works it cites.

fairseq: A Fast, Extensible Toolkit for Sequence Modeling

Ott, M.; Edunov, S.; Baevski, A.; Fan, A.; Gross, S.; Ng, N.; Grangier, D.; and Auli, M. 2019 · 2019

Later among the works it cites.

Neural cleanse: Identifying and mitigating backdoor attacks in neural networks

Wang, B.; Yao, Y.; Shan, S.; Li, H.; Viswanath, B.; Zheng, H.; and Zhao, B. Y. 2019 · 2019

Later among the works it cites.

Description based text classification with reinforcement learning

Chai, D.; Wu, W.; Han, Q.; Wu, F.; and Li, J. 2020 · 2020

Later among the works it cites.

Weight Poisoning Attacks on Pretrained Models

Kurita, K.; Michel, P.; and Neubig, G. 2020 · 2020

Later among the works it cites.

Best-First Beam Search

Meister, C.; Vieira, T.; and Cotterell, R. 2020 · 2020

Later among the works it cites.

Hidden trigger backdoor attacks

Saha, A.; Subramanya, A.; and Pirsiavash, H. 2020 · 2020

Later among the works it cites.

Pegasus: Pre-training with extracted gap-sentences for abstractive summarization

Zhang, J.; Zhao, Y.; Saleh, M.; and Liu, P. 2020 · 2020

Later among the works it cites.

Mitigating backdoor attacks in LSTM-based Text Classification Systems by Backdoor Keyword Identification

Chen, C.; and Dai, J. 2021 · 2021

Closest in time.

Triggerless Backdoor Attack for NLP Tasks with Clean Labels

Gan, L.; Li, J.; Zhang, T.; Li, X.; Meng, Y.; Wu, F.; Guo, S.; and Fan, C. 2021 · 2021

Closest in time.

Can Adversarial Weight Perturbations Inject Neural Backdoors

Garg, S.; Kumar, A.; Goel, V.; and Liang, Y. 2020 · 2032

Closest in time.

Be Careful about Poisoned Word Embeddings: Exploring the Vulnerability of the Embedding Layers in NLP Models

Yang, W.; Li, L.; Zhang, Z.; Ren, X.; Sun, X.; and He, B. 2021 · 2058

Closest in time.