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

This paper introduces a novel method to perform transfer learning acrossdomains and tasks, formulating it as a problem of learning to cluster. The keyinsight is that, in addition to features, we can transfer similarityinformation and this is sufficient to learn a similarity function andclustering network to perform both domain adaptation and cross-task transferlearning. We begin by reducing categorical information to pairwise constraints,which only considers whether two instances belong to the same class or not.This similarity is category-agnostic and can be learned from data in the sourcedomain using a similarity network. We then present two novel approaches forperforming transfer learning using this similarity function. First, forunsupervised domain adaptation, we design a new loss function to regularizeclassification with a constrained clustering loss, hence learning a clusteringnetwork with the transferred similarity metric generating the training inputs.Second, for cross-task learning (i.e., unsupervised clustering with unseencategories), we propose a framework to reconstruct and estimate the number ofsemantic clusters, again using the clustering network. Since the similaritynetwork is noisy, the key is to use a robust clustering algorithm, and we showthat our formulation is more robust than the alternative constrained andunconstrained clustering approaches. Using this method, we first show state ofthe art results for the challenging cross-task problem, applied on Omniglot andImageNet. Our results show that we can reconstruct semantic clusters with highaccuracy. We then evaluate the performance of cross-domain transfer usingimages from the Office-31 and SVHN-MNIST tasks and present top accuracy on bothdatasets. Our approach doesn't explicitly deal with domain discrepancy. If wecombine with a domain adaptation loss, it shows further improvement.

Learning to cluster in order to transfer across domains and tasks