Translational symmetry in convolutions with localized kernels causes an implicit bias toward high frequency adversarial examples.

This document provides a list of references to research papers and articles on topics related to neural networks, adversarial examples, image recognition, and deep learning. [Extracted from the article]