Understanding Robustness and Generalization of Artificial Neural Networks Through Fourier Masks
Name
frai-05-890016.pdf
Size
3.24 MB
Format
Adobe PDF
Checksum (MD5)
61c53dab76033c5164d36481b2319646
Author(s)
Karantzas, Nikos
Besier, Emma
Ortega Caro, Josue
Pitkow, Xaq
Tolias, Andreas S.
Patel, Ankit B.
Anselmi, Fabio
Date Issued
July 12, 2022
Journal
Frontiers in Artificial Intelligence
Publisher
Frontiers
Citation
Karantzas N, Besier E, Ortega Caro J, Pitkow X, Tolias AS, Patel AB and Anselmi F (2022) Understanding Robustness and Generalization of Artificial Neural Networks Through Fourier Masks. Front. Artif. Intell. 5:890016.
Version
Final published version
Abstract
Despite the enormous success of artificial neural networks (ANNs) in many disciplines, the characterization of their computations and the origin of key properties such as generalization and robustness remain open questions. Recent literature suggests that robust networks with good generalization properties tend to be biased toward processing low frequencies in images. To explore the frequency bias hypothesis further, we develop an algorithm that allows us to learn modulatory masks highlighting the essential input frequencies needed for preserving a trained network's performance. We achieve this by imposing invariance in the loss with respect to such modulations in the input frequencies. We first use our method to test the low-frequency preference hypothesis of adversarially trained or data-augmented networks. Our results suggest that adversarially robust networks indeed exhibit a low-frequency bias but we find this bias is also dependent on directions in frequency space. However, this is not necessarily true for other types of data augmentation. Our results also indicate that the essential frequencies in question are effectively the ones used to achieve generalization in the first place. Surprisingly, images seen through these modulatory masks are not recognizable and resemble texture-like patterns.
MIT Department
McGovern Institute for Brain Research at MIT
Lincoln Laboratory
Center for Brains, Minds, and Machines
Terms of Use
Creative Commons Attribution
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.3389/frai.2022.890016
