Timezone: »

 
ROPUST: Improving Robustness through Fine-tuning with Photonic Processors and Synthetic Gradients
Alessandro Cappelli · Ruben Ohana · Julien Launay · Laurent Meunier · Iacopo Poli

@
in A Blessing in Disguise: The Prospects and Perils of Adversarial Machine Learning »

Robustness to adversarial attacks is typically obtained through expensive adversarial training with Projected Gradient Descent. We introduce ROPUST, a remarkably simple and efficient method to leverage robust pre-trained models and further increase their robustness, at no cost in natural accuracy. Our technique relies on the use of an Optical Processing Unit (OPU), a photonic co-processor, and a fine-tuning step performed with Direct Feedback Alignment, a synthetic gradient training scheme. We test our method on nine different models against four attacks in RobustBench, consistently improving over state-of-the-art performance. We also introduce phase retrieval attacks, specifically designed to target our own defense. We show that even with state-of-the-art phase retrieval techniques, ROPUST is effective.

Author Information

Alessandro Cappelli (LightOn)
Ruben Ohana (Ecole Normale Supérieure & LightOn)
Julien Launay (École Normale Supérieure)
Laurent Meunier (Paris Dauphine University / Facebook AI Research)
Iacopo Poli (LightOn)

More from the Same Authors