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

Contrastive representation learning has shown to be effective to learnrepresentations from unlabeled data. However, much progress has been made invision domains relying on data augmentations carefully designed using domainknowledge. In this work, we propose i-Mix, a simple yet effectivedomain-agnostic regularization strategy for improving contrastiverepresentation learning. We cast contrastive learning as training anon-parametric classifier by assigning a unique virtual class to each data in abatch. Then, data instances are mixed in both the input and virtual labelspaces, providing more augmented data during training. In experiments, wedemonstrate that i-Mix consistently improves the quality of learnedrepresentations across domains, including image, speech, and tabular data.Furthermore, we confirm its regularization effect via extensive ablationstudies across model and dataset sizes. The code is available athttps://github.com/kibok90/imix.

i-Mix: A Domain-Agnostic Strategy for Contrastive Representation Learning