Connecting Lyapunov Control Theory to Adversarial Attacks
This work addresses the challenge of improving security for neural network systems, though it is incremental as it focuses on a weaker adversary.
The paper tackles the problem of defending neural networks against adversarial attacks by leveraging control theory, resulting in a provable defense against a weaker adversary to demonstrate the approach's potential.
Significant work is being done to develop the math and tools necessary to build provable defenses, or at least bounds, against adversarial attacks of neural networks. In this work, we argue that tools from control theory could be leveraged to aid in defending against such attacks. We do this by example, building a provable defense against a weaker adversary. This is done so we can focus on the mechanisms of control theory, and illuminate its intrinsic value.