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

Deep metric learning plays a key role in various machine learning tasks. Mostof the previous works have been confined to sampling from a mini-batch, whichcannot precisely characterize the global geometry of the embedding space.Although researchers have developed proxy- and classification-based methods totackle the sampling issue, those methods inevitably incur a redundantcomputational cost. In this paper, we propose a novel Proxy-based deep GraphMetric Learning (ProxyGML) approach from the perspective of graphclassification, which uses fewer proxies yet achieves better comprehensiveperformance. Specifically, multiple global proxies are leveraged tocollectively approximate the original data points for each class. Toefficiently capture local neighbor relationships, a small number of suchproxies are adaptively selected to construct similarity subgraphs between theseproxies and each data point. Further, we design a novel reverse labelpropagation algorithm, by which the neighbor relationships are adjustedaccording to ground-truth labels, so that a discriminative metric space can belearned during the process of subgraph classification. Extensive experimentscarried out on widely-used CUB-200-2011, Cars196, and Stanford Online Productsdatasets demonstrate the superiority of the proposed ProxyGML over thestate-of-the-art methods in terms of both effectiveness and efficiency. Thesource code is publicly available at https://github.com/YuehuaZhu/ProxyGML.

Fewer is More: A Deep Graph Metric Learning Perspective Using Fewer Proxies