2019

ME-Net: Towards Effective Adversarial Robustness with Matrix Estimation

Yang, Yuzhe, Zhang, Guo, Katabi, Dina et al.

Understand

Deep neural networks are vulnerable to adversarial attacks.

The literature is rich with algorithms that can easily craft successful adversarial examples.
In contrast, the performance of defense techniques still lags behind.
This paper proposes ME-Net, a defense method that leverages matrix estimation (ME).

Built on

Lessons from the netflix prize challenge
Bell, R. M. and Koren, Y · 1931
Earlier work this paper cites.
Visualizing data using t-sne
Maaten, L. v. d. and Hinton, G · 2008
Earlier work this paper cites.
Exact matrix completion via convex optimization
Candès, E. J. and Recht, B · 2009
Earlier work this paper cites.
Matrix completion from noisy entries
Keshavan, R. H., Montanari, A., and Oh, S · 2010
Earlier work this paper cites.
Spectral regularization algorithms for learning large incomplete matrices
Mazumder, R., Hastie, T., and Tibshirani, R · 2010
Earlier work this paper cites.
Stochastic blockmodel approximation of a graphon: Theory and consistent estimation
Airoldi, E. M., Costa, T. B., and Chan, S. H · 2013
Earlier work this paper cites.
Low-rank matrix completion using alternating minimization
Jain, P., Netrapalli, P., and Sanghavi, S · 2013
Earlier work this paper cites.
Intriguing properties of neural networks
Original
Szegedy, C., Zaremba, W., Sutskever, I., Bruna, J., Erhan, D., Goodfellow, I., and Fergus, R · 2013
Earlier work this paper cites.
Community detection in general stochastic block models: Fundamental limits and efficient algorithms for recovery
Abbe, E. and Sandon, C · 2015
Earlier work this paper cites.
Matrix estimation by universal singular value thresholding
Chatterjee, S. et al · 2015
Earlier work this paper cites.
Fast low-rank estimation by projected gradient descent: General statistical and algorithmic guarantees
Original
Chen, Y. and Wainwright, M. J · 2015
Earlier work this paper cites.
Explaining and harnessing adversarial examples
Original
Goodfellow, I., Shlens, J., and Szegedy, C · 2015
Earlier work this paper cites.

Similar

An overview of low-rank matrix recovery from incomplete observations
Original
Davenport, M. A. and Romberg, J · 2016
Cited alongside, same era.
Matrix completion has no spurious local minimum
Ge, R., Lee, J. D., and Ma, T · 2016
Cited alongside, same era.
Deep residual learning for image recognition
He, K., Zhang, X., Ren, S., and Sun, J · 2016
Cited alongside, same era.
Thy friend is my friend: Iterative collaborative filtering for sparse matrix estimation
Borgs, C., Chayes, J., Lee, C. E., and Shah, D · 2017
Cited alongside, same era.
Decision-based adversarial attacks: Reliable attacks against black-box machine learning models
Original
Brendel, W., Rauber, J., and Bethge, M · 2017
Cited alongside, same era.
Towards evaluating the robustness of neural networks
Carlini, N. and Wagner, D · 2017
Cited alongside, same era.
Countering adversarial images using input transformations
Original
Guo, C., Rana, M., Cisse, M., and van der Maaten, L · 2017
Cited alongside, same era.
Densely connected convolutional networks
Huang, G., Liu, Z., van der Maaten, L., and Weinberger, K. Q · 2017
Cited alongside, same era.
Towards deep learning models resistant to adversarial attacks
Original
Madry, A., Makelov, A., Schmidt, L., Tsipras, D., and Vladu, A · 2017
Cited alongside, same era.
Feature squeezing: Detecting adversarial examples in deep neural networks
Original
Xu, W., Evans, D., and Qi, Y · 2017
Cited alongside, same era.
Model agnostic time series analysis via matrix estimation
Agarwal, A., Amjad, M. J., Shah, D., and Shen, D · 2018
Cited alongside, same era.
Threat of adversarial attacks on deep learning in computer vision: A survey
Original
Akhtar, N. and Mian, A · 2018
Cited alongside, same era.

Then

Obfuscated gradients give a false sense of security: Circumventing defenses to adversarial examples
Original
Athalye, A., Carlini, N., and Wagner, D · 2018
Later among the works it cites.
Thermometer encoding: One hot way to resist adversarial examples
Buckman, J., Roy, A., Raffel, C., and Goodfellow, I · 2018
Later among the works it cites.
Improving adversarial robustness by data-specific discretization
Original
Chen, J., Wu, X., Liang, Y., and Jha, S · 2018
Later among the works it cites.
Harnessing structures in big data via guaranteed low-rank matrix estimation
Original
Chen, Y. and Chi, Y · 2018
Later among the works it cites.
Certified robustness to adversarial examples with differential privacy
Original
Lecuyer, M., Atlidakis, V., Geambasu, R., Hsu, D., and Jana, S · 2018
Later among the works it cites.
Logit pairing methods can fool gradient-based attacks
Mosbach, M., Andriushchenko, M., Trost, T., Hein, M., and Klakow, D · 2018
Later among the works it cites.
Defense-gan: Protecting classifiers against adversarial attacks using generative models
Samangouei, P., Kabkab, M., and Chellappa, R · 2018
Later among the works it cites.
Adversarially robust generalization requires more data
Original
Schmidt, L., Santurkar, S., Tsipras, D., Talwar, K., and Madry, A · 2018
Later among the works it cites.
Pixeldefend: Leveraging generative models to understand and defend against adversarial examples
Song, Y., Kim, T., Nowozin, S., Ermon, S., and Kushman, N · 2018
Later among the works it cites.
Adversarial risk and the dangers of evaluating against weak attacks
Original
Uesato, J., O’Donoghue, B., Oord, A. v. d., and Kohli, P · 2018
Later among the works it cites.
Mitigating adversarial effects through randomization
Xie, C., Wang, J., Zhang, Z., Ren, Z., and Yuille, A · 2018
Later among the works it cites.

Beyond the bibliography

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

Open on alphaXiv

alphaXiv is searching for related work…

Reading the bibliography…