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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Novelty detection is the process of determining whether a query examplediffers from the learned training distribution. Previous methods attempt tolearn the representation of the normal samples via generative adversarialnetworks (GANs). However, they will suffer from instability training, modedropping, and low discriminative ability. Recently, various pretext tasks (e.g.rotation prediction and clustering) have been proposed for self-supervisedlearning in novelty detection. However, the learned latent features are stilllow discriminative. We overcome such problems by introducing a noveldecoder-encoder framework. Firstly, a generative network (a.k.a. decoder)learns the representation by mapping the initialized latent vector to an image.In particular, this vector is initialized by considering the entiredistribution of training data to avoid the problem of mode-dropping. Secondly,a contrastive network (a.k.a. encoder) aims to ``learn to compare'' throughmutual information estimation, which directly helps the generative network toobtain a more discriminative representation by using a negative dataaugmentation strategy. Extensive experiments show that our model hassignificant superiority over cutting-edge novelty detectors and achieves newstate-of-the-art results on some novelty detection benchmarks, e.g. CIFAR10 andDCASE. Moreover, our model is more stable for training in a non-adversarialmanner, compared to other adversarial based novelty detection methods.

Novelty Detection via Contrastive Learning with Negative Data Augmentation