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 (DNNs) have achieved tremendous success in a variety ofapplications across many disciplines. Yet, their superior performance comeswith the expensive cost of requiring correctly annotated large-scale datasets.Moreover, due to DNNs' rich capacity, errors in training labels can hamperperformance. To combat this problem, mean absolute error (MAE) has recentlybeen proposed as a noise-robust alternative to the commonly-used categoricalcross entropy (CCE) loss. However, as we show in this paper, MAE can performpoorly with DNNs and challenging datasets. Here, we present a theoreticallygrounded set of noise-robust loss functions that can be seen as ageneralization of MAE and CCE. Proposed loss functions can be readily appliedwith any existing DNN architecture and algorithm, while yielding goodperformance in a wide range of noisy label scenarios. We report results fromexperiments conducted with CIFAR-10, CIFAR-100 and FASHION-MNIST datasets andsynthetically generated noisy labels.

Generalized Cross Entropy Loss for Training Deep Neural Networks with Noisy Labels