A geometric statistic for deep learning model confidence and adversarial defense

Donald Waagen; Don Hulsey; Jamie Godwin; David Gray

doi:10.1117/12.2618299

31 May 2022 A geometric statistic for deep learning model confidence and adversarial defense

Donald Waagen, Don Hulsey, Jamie Godwin, David Gray

Proceedings Volume 12096, Automatic Target Recognition XXXII; 1209606 (2022) https://doi.org/10.1117/12.2618299
Event: SPIE Defense + Commercial Sensing, 2022, Orlando, Florida, United States

Abstract

Deep learning models are pervasive for a multitude of tasks, but the complexity of these models can limit interpretation and inhibit trust in their estimates of confidence. For the classification task, we investigate the induced geometric relationships between the class conditioned data distributions with the deep learning models’ output weight vectors. We propose a simple statistic, which we call Angular Margin, to characterize the “confidence” of the model given a new input. We compare and contrast our statistic to Angular Visual Hardness and Softmax outputs. We demonstrate that Angular Margin provides a superior statistic for detecting minimum-perturbation adversarial attacks and/or misclassified images than standard Softmax predictions.

Conference Presentation

Citation Download Citation

Donald Waagen, Don Hulsey, Jamie Godwin, and David Gray "A geometric statistic for deep learning model confidence and adversarial defense", Proc. SPIE 12096, Automatic Target Recognition XXXII, 1209606 (31 May 2022); https://doi.org/10.1117/12.2618299

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available