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

Network representation learning (NRL) technique has been successfully adoptedin various data mining and machine learning applications. Random walk based NRLis one popular paradigm, which uses a set of random walks to capture thenetwork structural information, and then employs word2vec models to learn thelow-dimensional representations. However, until now there is lack of aframework, which unifies existing random walk based NRL models and supports toefficiently learn from large networks. The main obstacle comes from the diverserandom walk models and the inefficient sampling method for the random walkgeneration. In this paper, we first introduce a new and efficient edge samplerbased on Metropolis-Hastings sampling technique, and theoretically show theconvergence property of the edge sampler to arbitrary discrete probabilitydistributions. Then we propose a random walk model abstraction, in which userscan easily define different transition probability by specifying dynamic edgeweights and random walk states. The abstraction is efficiently supported by ouredge sampler, since our sampler can draw samples from unnormalized probabilitydistribution in constant time complexity. Finally, with the new edge samplerand random walk model abstraction, we carefully implement a scalable NRLframework called UniNet. We conduct comprehensive experiments with five randomwalk based NRL models over eleven real-world datasets, and the results clearlydemonstrate the efficiency of UniNet over billion-edge networks.

UniNet: Scalable Network Representation Learning with Metropolis-Hastings Sampling