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

It is essential for safety-critical applications of deep neural networks todetermine when new inputs are significantly different from the trainingdistribution. In this paper, we explore this out-of-distribution (OOD)detection problem for image classification using clusters of semanticallysimilar embeddings of the training data and exploit the differences in distancerelationships to these clusters between in- and out-of-distribution data. Westudy the structure and separation of clusters in the embedding space and findthat supervised contrastive learning leads to well-separated clusters while itsself-supervised counterpart fails to do so. In our extensive analysis ofdifferent training methods, clustering strategies, distance metrics, andthresholding approaches, we observe that there is no clear winner. The optimalapproach depends on the model architecture and selected datasets for in- andout-of-distribution. While we could reproduce the outstanding results forcontrastive training on CIFAR-10 as in-distribution data, we find standardcross-entropy paired with cosine similarity outperforms all contrastivetraining methods when training on CIFAR-100 instead. Cross-entropy providescompetitive results as compared to expensive contrastive training methods.

Is it all a cluster game? -- Exploring Out-of-Distribution Detection based on Clustering in the Embedding Space