LGCRMLJan 2, 2020

ATHENA: A Framework based on Diverse Weak Defenses for Building Adversarial Defense

arXiv:2001.00308v27 citations
AI Analysis

This addresses the need for generic and extensible adversarial defenses, which is an incremental improvement over domain-specific methods.

The paper tackles the problem of adversarial attacks by proposing ATHENA, a flexible framework based on diverse weak defenses, and shows it is effective across multiple threat models including zero-knowledge, black-box, gray-box, and white-box attacks.

There has been extensive research on developing defense techniques against adversarial attacks; however, they have been mainly designed for specific model families or application domains, therefore, they cannot be easily extended. Based on the design philosophy of ensemble of diverse weak defenses, we propose ATHENA---a flexible and extensible framework for building generic yet effective defenses against adversarial attacks. We have conducted a comprehensive empirical study to evaluate several realizations of ATHENA with four threat models including zero-knowledge, black-box, gray-box, and white-box. We also explain (i) why diversity matters, (ii) the generality of the defense framework, and (iii) the overhead costs incurred by ATHENA.

Code Implementations3 repos
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes