Mind the Pad -- CNNs Can Develop Blind Spots

Bilal Alsallakh, Narine Kokhlikyan, Vivek Miglani, Jun Yuan, Orion Reblitz-Richardson

Keywords: CNN, convolution, spatial bias, blind spots, foveation, padding, exposition, debugging, visualization

Abstract: We show how feature maps in convolutional networks are susceptible to spatial bias. Due to a combination of architectural choices, the activation at certain locations is systematically elevated or weakened. The major source of this bias is the padding mechanism. Depending on several aspects of convolution arithmetic, this mechanism can apply the padding unevenly, leading to asymmetries in the learned weights. We demonstrate how such bias can be detrimental to certain tasks such as small object detection: the activation is suppressed if the stimulus lies in the impacted area, leading to blind spots and misdetection. We explore alternative padding methods and propose solutions for analyzing and mitigating spatial bias.

One-sentence Summary: The padding mechanism in CNNs can induce harmful spatial bias in the learned weights and in the feature maps, which can be mitigated with careful architectural choices.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Community Implementations: [ 2 code implementations](https://www.catalyzex.com/paper/arxiv:2010.02178/code)