我还能再战! (`-`)
我还能再战! (`-`)
BERT-ATTACK: Adversarial Attack Against BERT Using BERT Dependencies Python 3.7 PyTorch1.4.0 transformers2.9.0 TextFooler Usage To train a classification model, please use the run_glue.py script in the huggingface transformers==2.9.0. To generate adversarial samples based on the masked-LM, run ...
We show that it is possible to improve the accuracy of BERT-based models in the face of adversarial attacks without significantly reducing the accuracy for regular benchmark samples. We present co-tuning with a synthetic data generator as a highly effective method to protect against 95% of pre...
BERT-ATTACK: Adversarial attack against BERT using BERT. In Proceed- ings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 6193–6202, Online, November 2020. Association for Computational Linguistics. 39. Lifan Yuan, YiChi Zhang, Yangyi Chen, and Wei Wei...
Bert-attack: Adversarial attack against bert using bert. arXiv preprint arXiv:2004.09984. Google Scholar Liang et al., 2017 Liang, B., Li, H., Su, M., Bian, P., Li, X., Shi, W., 2017. Deep text classification can be fooled. arXiv preprint arXiv:1704.08006. Google Scholar Liu ...
bert-attackUntargeted ClassificationUSE sentence encoding cosine similarity, Maximum number of words perturbedBERT Masked Token Prediction (with subword expansion)Greedy-WIR(["BERT-ATTACK: Adversarial Attack Against BERT Using BERT" (Li et al., 2020)](https://arxiv.org/abs/2004.09984)) ...
FireBERT: Hardening BERT Classif i ers AgainstAdversarial AttackGunnar Mein, Kevin Hartman, Andrew MorrisUC Berkeley MIDS studentsgunnarmein@berkeley.edu, kevin.hartman@berkeley.edu, andrewmorris@berkeley.eduAbstractWe present FireBERT, a set of three proof-of-concept NLP classif i ers hardenedagains...
X. (2020). BERT-ATTACK: adversarial attack against BERT using BERT. In B. Webber, T. Cohn, Y. He, et al. (Eds.), Proceedings of the 2020 conference on empirical methods in natural language processing (pp. 6193–6202). Stroudsburg: ACL. Google Scholar Lu, D., Wang, Z., Wang,...
Li, L., Ma, R., Guo, Q., Xue, X., Qiu, X.: BERT-ATTACK: Adversarial attack against BERT using BERT. In: Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP). ACL, November 2020.https://doi.org/10.18653/v1/2020.emnlp-main.500 ...