Attention Meets Perturbations: Robust and Interpretable Attention with Adversarial Training
@article{Kitada2020AttentionMP, title={Attention Meets Perturbations: Robust and Interpretable Attention with Adversarial Training}, author={Shunsuke Kitada and H. Iyatomi}, journal={ArXiv}, year={2020}, volume={abs/2009.12064} }
In recent years, deep learning models have placed more emphasis on the interpretability and robustness of models. The attention mechanism is an important technique that contributes to these elements and is widely used, especially in the natural language processing (NLP) field. Adversarial training (AT) is a powerful regularization technique for enhancing the robustness of neural networks and has been successful in many applications. The application of AT to the attention mechanism is expected… CONTINUE READING
References
SHOWING 1-10 OF 45 REFERENCES
Robust Multilingual Part-of-Speech Tagging via Adversarial Training
- Computer Science
- NAACL-HLT
- 2018
- 51
- PDF
Interpretable Adversarial Perturbation in Input Embedding Space for Text
- Computer Science, Mathematics
- IJCAI
- 2018
- 59
- Highly Influential
- PDF
Explaining and Harnessing Adversarial Examples
- Computer Science, Mathematics
- ICLR
- 2015
- 6,252
- Highly Influential
- PDF
Adversarial Training Methods for Semi-Supervised Text Classification
- Computer Science, Mathematics
- ICLR
- 2017
- 337
- Highly Influential
- PDF
A Theoretical Framework for Robustness of (Deep) Classifiers against Adversarial Samples
- Computer Science, Mathematics
- ICLR
- 2017
- 36
- PDF
Understanding adversarial training: Increasing local stability of supervised models through robust optimization
- Computer Science, Mathematics
- Neurocomputing
- 2018
- 127
- PDF