Irina Nicolae, PhD, is currently a research scientist in the Cognitive Datascience team of IBM Research, Dublin. Her main interests include learning representations for complex data and security for deployed models. She has received her PhD from University of Saint-Etienne, France, for a research project on similarity learning with theoretical guarantees for numerical and temporal data. Previously, she has graduated from Politehnica University of Bucharest in Computer Science in 2011, and from ENSIMAG in Information Systems in 2013.

Efficient Defenses Against Adversarial Examples for Deep Neural Networks

Following the recent adoption of deep neural networks (DNN) in a wide range of application fields, adversarial attacks against these models have proven to be an indisputable threat. Adversarial samples are crafted with a deliberate intention of producing a specific response from the system. Multiple attacks and defenses have been proposed in the literature, but the lack of better understanding of sensitivity of DNNs justifies adversarial samples still being an open question.

In this talk, we propose a new defense method based on practical observations which is easy to integrate into models and performs better than state-of-the-art defenses. Our proposed solution is meant to reinforce the structure of a DNN, making its prediction more stable and less likely to be fooled by adversarial samples.

The extensive experimental study proves the efficiency of our method against multiple attacks, comparing it to multiple defenses, both in white-box and black-box setups. Additionally, the implementation brings almost no overhead to the training procedure, while maintaining the prediction performance of the original model on clean samples. A live demo of creating adversarial images will take place during the talk.