The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
CPU empat teras telah menjadi konfigurasi desktop biasa untuk pejabat hari ini. Peningkatan bilangan pemproses pada satu cip membuka peluang baharu untuk pengkomputeran selari. Matlamat kami adalah untuk menggunakan seni bina berbilang teras serta berbilang pemproses untuk mempercepatkan algoritma perlombongan data berskala besar. Dalam kertas kerja ini, kami membentangkan rangka kerja pembelajaran selari am, Potong-Dan-Jahit, untuk melatih model rantai Markov tersembunyi. Khususnya, kami mencadangkan dua varian khusus model, CAS-LDS untuk pembelajaran sistem dinamik linear (LDS) dan CAS-HMM untuk mempelajari model Markov tersembunyi (HMM). Sumbangan utama kami ialah kaedah baru untuk mengendalikan kebergantungan data disebabkan oleh struktur rantai pembolehubah tersembunyi, untuk menyelaraskan algoritma pembelajaran parameter berasaskan EM. Kami melaksanakan CAS-LDS dan CAS-HMM menggunakan OpenMP pada dua superkomputer dan desktop komersial empat teras. Keputusan eksperimen menunjukkan bahawa algoritma selari menggunakan Potong-Dan-Jahit mencapai ketepatan yang setanding dan kelajuan hampir linear berbanding versi bersiri tradisional.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Salinan
Lei LI, Bin FU, Christos FALOUTSOS, "Efficient Parallel Learning of Hidden Markov Chain Models on SMPs" in IEICE TRANSACTIONS on Information,
vol. E93-D, no. 6, pp. 1330-1342, June 2010, doi: 10.1587/transinf.E93.D.1330.
Abstract: Quad-core cpus have been a common desktop configuration for today's office. The increasing number of processors on a single chip opens new opportunity for parallel computing. Our goal is to make use of the multi-core as well as multi-processor architectures to speed up large-scale data mining algorithms. In this paper, we present a general parallel learning framework, Cut-And-Stitch, for training hidden Markov chain models. Particularly, we propose two model-specific variants, CAS-LDS for learning linear dynamical systems (LDS) and CAS-HMM for learning hidden Markov models (HMM). Our main contribution is a novel method to handle the data dependencies due to the chain structure of hidden variables, so as to parallelize the EM-based parameter learning algorithm. We implement CAS-LDS and CAS-HMM using OpenMP on two supercomputers and a quad-core commercial desktop. The experimental results show that parallel algorithms using Cut-And-Stitch achieve comparable accuracy and almost linear speedups over the traditional serial version.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E93.D.1330/_p
Salinan
@ARTICLE{e93-d_6_1330,
author={Lei LI, Bin FU, Christos FALOUTSOS, },
journal={IEICE TRANSACTIONS on Information},
title={Efficient Parallel Learning of Hidden Markov Chain Models on SMPs},
year={2010},
volume={E93-D},
number={6},
pages={1330-1342},
abstract={Quad-core cpus have been a common desktop configuration for today's office. The increasing number of processors on a single chip opens new opportunity for parallel computing. Our goal is to make use of the multi-core as well as multi-processor architectures to speed up large-scale data mining algorithms. In this paper, we present a general parallel learning framework, Cut-And-Stitch, for training hidden Markov chain models. Particularly, we propose two model-specific variants, CAS-LDS for learning linear dynamical systems (LDS) and CAS-HMM for learning hidden Markov models (HMM). Our main contribution is a novel method to handle the data dependencies due to the chain structure of hidden variables, so as to parallelize the EM-based parameter learning algorithm. We implement CAS-LDS and CAS-HMM using OpenMP on two supercomputers and a quad-core commercial desktop. The experimental results show that parallel algorithms using Cut-And-Stitch achieve comparable accuracy and almost linear speedups over the traditional serial version.},
keywords={},
doi={10.1587/transinf.E93.D.1330},
ISSN={1745-1361},
month={June},}
Salinan
TY - JOUR
TI - Efficient Parallel Learning of Hidden Markov Chain Models on SMPs
T2 - IEICE TRANSACTIONS on Information
SP - 1330
EP - 1342
AU - Lei LI
AU - Bin FU
AU - Christos FALOUTSOS
PY - 2010
DO - 10.1587/transinf.E93.D.1330
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E93-D
IS - 6
JA - IEICE TRANSACTIONS on Information
Y1 - June 2010
AB - Quad-core cpus have been a common desktop configuration for today's office. The increasing number of processors on a single chip opens new opportunity for parallel computing. Our goal is to make use of the multi-core as well as multi-processor architectures to speed up large-scale data mining algorithms. In this paper, we present a general parallel learning framework, Cut-And-Stitch, for training hidden Markov chain models. Particularly, we propose two model-specific variants, CAS-LDS for learning linear dynamical systems (LDS) and CAS-HMM for learning hidden Markov models (HMM). Our main contribution is a novel method to handle the data dependencies due to the chain structure of hidden variables, so as to parallelize the EM-based parameter learning algorithm. We implement CAS-LDS and CAS-HMM using OpenMP on two supercomputers and a quad-core commercial desktop. The experimental results show that parallel algorithms using Cut-And-Stitch achieve comparable accuracy and almost linear speedups over the traditional serial version.
ER -