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

Large-scale supervised classification algorithms, especially those based ondeep convolutional neural networks (DCNNs), require vast amounts of trainingdata to achieve state-of-the-art performance. Decreasing this data requirementwould significantly speed up the training process and possibly improvegeneralization. Motivated by this objective, we consider the task of adaptivelyfinding concise training subsets which will be iteratively presented to thelearner. We use convex optimization methods, based on an objective criterionand feedback from the current performance of the classifier, to efficientlyidentify informative samples to train on. We propose an algorithm to decomposethe optimization problem into smaller per-class problems, which can be solvedin parallel. We test our approach on standard classification tasks anddemonstrate its effectiveness in decreasing the training set size withoutcompromising performance. We also show that our approach can make theclassifier more robust in the presence of label noise and class imbalance.

DCNNs on a Diet: Sampling Strategies for Reducing the Training Set Size