English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Deep neural networks have been used in various fields, but their internalbehavior is not well known. In this study, we discuss two counterintuitivebehaviors of convolutional neural networks (CNNs). First, we evaluated the sizeof the receptive field. Previous studies have attempted to increase or controlthe size of the receptive field. However, we observed that the size of thereceptive field does not describe the classification accuracy. The size of thereceptive field would be inappropriate for representing superiority inperformance because it reflects only depth or kernel size and does not reflectother factors such as width or cardinality. Second, using the effectivereceptive field, we examined the pixels contributing to the output.Intuitively, each pixel is expected to equally contribute to the final output.However, we found that there exist pixels in a partially dead state with littlecontribution to the output. We reveal that the reason for this lies in thearchitecture of CNN and discuss solutions to reduce the phenomenon.Interestingly, for general classification tasks, the existence of dead pixelsimproves the training of CNNs. However, in a task that captures smallperturbation, dead pixels degrade the performance. Therefore, the existence ofthese dead pixels should be understood and considered in practical applicationsof CNN.

Dead Pixel Test Using Effective Receptive Field