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

We address the problem of network calibration adjusting miscalibratedconfidences of deep neural networks. Many approaches to network calibrationadopt a regularization-based method that exploits a regularization term tosmooth the miscalibrated confidences. Although these approaches have shown theeffectiveness on calibrating the networks, there is still a lack ofunderstanding on the underlying principles of regularization in terms ofnetwork calibration. We present in this paper an in-depth analysis of existingregularization-based methods, providing a better understanding on how theyaffect to network calibration. Specifically, we have observed that 1) theregularization-based methods can be interpreted as variants of label smoothing,and 2) they do not always behave desirably. Based on the analysis, we introducea novel loss function, dubbed ACLS, that unifies the merits of existingregularization methods, while avoiding the limitations. We show extensiveexperimental results for image classification and semantic segmentation onstandard benchmarks, including CIFAR10, Tiny-ImageNet, ImageNet, and PASCALVOC, demonstrating the effectiveness of our loss function.

ACLS: Adaptive and Conditional Label Smoothing for Network Calibration