Adversarial Robustness in Parameter-Space Classifiers
This addresses the reliability and applicability issues of machine learning methods in adversarial settings, offering a potential solution for domains where security is critical, though it appears incremental as it builds on existing INR frameworks.
The paper tackles the problem of adversarial vulnerability in machine learning by showing that parameter-space classifiers based on Implicit Neural Representations (INRs) are inherently robust to adversarial attacks without requiring robust training, as demonstrated through a novel suite of attacks and analysis.
Implicit Neural Representations (INRs) have been recently garnering increasing interest in various research fields, mainly due to their ability to represent large, complex data in a compact and continuous manner. Past work further showed that numerous popular downstream tasks can be performed directly in the INR parameter-space. Doing so can substantially reduce the computational resources required to process the represented data in their native domain. A major difficulty in using modern machine-learning approaches, is their high susceptibility to adversarial attacks, which have been shown to greatly limit the reliability and applicability of such methods in a wide range of settings. In this work, we show that parameter-space models trained for classification are inherently robust to adversarial attacks -- without the need of any robust training. To support our claims, we develop a novel suite of adversarial attacks targeting parameter-space classifiers, and furthermore analyze practical considerations of attacking parameter-space classifiers.