caa a22 4500 60546846X CHVBK 20210128100316.0 cr unu---uuuuu 210128e20150101xx s 000 0 eng 10.1007/s00500-014-1376-8 doi (NATIONALLICENCE)springer-10.1007/s00500-014-1376-8 Communication-efficient algorithms for parallel latent Dirichlet allocation [Elektronische Daten] [Jian-Feng Yan, Jia Zeng, Yang Gao, Zhi-Qiang Liu] Latent Dirichlet allocation (LDA) is a popular topic modeling method which has found many multimedia applications, such as motion analysis and image categorization. Communication cost is one of the main bottlenecks for large-scale parallel learning of LDA. To reduce communication cost, we introduce Zipf's law and propose novel parallel LDA algorithms that communicate only partial important information at each learning iteration. The proposed algorithms are much more efficient than the current state-of-the-art algorithms in both communication and computation costs. Extensive experiments on large-scale data sets demonstrate that our algorithms can greatly reduce communication and computation costs to achieve a better scalability. Springer-Verlag Berlin Heidelberg, 2014 Latent Dirichlet allocation nationallicence Parallel learning nationallicence Zipf's law nationallicence Belief propagation nationallicence Gibbs sampling nationallicence Yan Jian-Feng School of Computer Science and Technology, Soochow University, 215006, Suzhou, China aut Zeng Jia School of Computer Science and Technology, Soochow University, 215006, Suzhou, China aut Gao Yang School of Computer Science and Technology, Soochow University, 215006, Suzhou, China aut Liu Zhi-Qiang School of Creative Media, City University of Hong Kong, Hong Kong, China aut Soft Computing Springer Berlin Heidelberg 19/1(2015-01-01), 3-11 1432-7643 19:1<3 2015 19 500 https://doi.org/10.1007/s00500-014-1376-8 text/html Onlinezugriff via DOI BK010053 XK010053 XK010000 Metadata rights reserved Springer special CC-BY-NC licence nationallicence 1 research-article jats NATIONALLICENCE NATIONALLICENCE NL-springer NATIONALLICENCE 856 40 https://doi.org/10.1007/s00500-014-1376-8 text/html Onlinezugriff via DOI NATIONALLICENCE 700 1- Yan Jian-Feng School of Computer Science and Technology, Soochow University, 215006, Suzhou, China aut NATIONALLICENCE 700 1- Zeng Jia School of Computer Science and Technology, Soochow University, 215006, Suzhou, China aut NATIONALLICENCE 700 1- Gao Yang School of Computer Science and Technology, Soochow University, 215006, Suzhou, China aut NATIONALLICENCE 700 1- Liu Zhi-Qiang School of Creative Media, City University of Hong Kong, Hong Kong, China aut NATIONALLICENCE 773 0- Soft Computing Springer Berlin Heidelberg 19/1(2015-01-01), 3-11 1432-7643 19:1<3 2015 19 500