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

Influence maximization aims to select k most-influential vertices or seeds ina network, where influence is defined by a given diffusion process. Althoughcomputing optimal seed set is NP-Hard, efficient approximation algorithmsexist. However, even state-of-the-art parallel implementations are limited by asampling step that incurs large memory footprints. This in turn limits theproblem size reach and approximation quality. In this work, we study the memoryfootprint of the sampling process collecting reverse reachability informationin the IMM (Influence Maximization via Martingales) algorithm over largereal-world social networks. We present a memory-efficient optimization approach(called HBMax) based on Ripples, a state-of-the-art multi-threaded parallelinfluence maximization solution. Our approach, HBMax, uses a portion of thereverse reachable (RR) sets collected by the algorithm to learn thecharacteristics of the graph. Then, it compresses the intermediate reversereachability information with Huffman coding or bitmap coding, and queries onthe partially decoded data, or directly on the compressed data to preserve thememory savings obtained through compression. Considering a NUMA architecture,we scale up our solution on 64 CPU cores and reduce the memory footprint by upto 82.1% with average 6.3% speedup (encoding overhead is offset by performancegain from memory reduction) without loss of accuracy. For the largest testedgraph Twitter7 (with 1.4 billion edges), HBMax achieves 5.9X compression ratioand 2.2X speedup.

HBMax: Optimizing Memory Efficiency for Parallel Influence Maximization on Multicore Architectures