Dimension Reduction Methods for Collaborative Mobile Gossip Learning

Arpad Berta, Istvan Hegedus, M. Jelasity

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

Decentralized learning algorithms are very sensitive to the size of the raw data records due to the resulting large communication cost. This can, in the worst case, even make decentralized learning infeasible. Dimension reduction is a key technique to compress data and to obtain small models. In this paper, we propose a number of robust and efficient decentralized approaches to dimension reduction in the system model where each network node holds only one data record. These algorithms build on searching for good random projections. We present a thorough experimental comparison of the proposed algorithms and compare them with a variant of distributed singular value decomposition (SVD), a state-of-the-art algorithm for dimension reduction. We base our experiments on a trace of real mobile phone usage. We conclude that our method based on selecting good random projections is preferable and provides good quality results when the output is required on a very short timescale, within tens of minutes. We also present a hybrid method that combines the advantages of random projections and SVD. We demonstrate that the hybrid method offers good performance over all timescales.

Original languageEnglish
Title of host publicationProceedings - 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages393-397
Number of pages5
ISBN (Print)9781467387750
DOIs
Publication statusPublished - Mar 31 2016
Event24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2016 - Heraklion, Crete, Greece
Duration: Feb 17 2016Feb 19 2016

Other

Other24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2016
CountryGreece
CityHeraklion, Crete
Period2/17/162/19/16

Fingerprint

Random Projection
Gossip
Dimension Reduction
Reduction Method
Decentralized
Singular value decomposition
Hybrid Method
Time Scales
Communication Cost
Mobile Phone
Mobile phones
Learning algorithms
Learning Algorithm
Trace
Output
Communication
Vertex of a graph
Model
Demonstrate
Experiment

Keywords

  • dimension reduction
  • distributed data mining
  • gossip

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Hardware and Architecture
  • Software
  • Control and Optimization

Cite this

Berta, A., Hegedus, I., & Jelasity, M. (2016). Dimension Reduction Methods for Collaborative Mobile Gossip Learning. In Proceedings - 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2016 (pp. 393-397). [7445364] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/PDP.2016.20

Dimension Reduction Methods for Collaborative Mobile Gossip Learning. / Berta, Arpad; Hegedus, Istvan; Jelasity, M.

Proceedings - 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2016. Institute of Electrical and Electronics Engineers Inc., 2016. p. 393-397 7445364.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Berta, A, Hegedus, I & Jelasity, M 2016, Dimension Reduction Methods for Collaborative Mobile Gossip Learning. in Proceedings - 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2016., 7445364, Institute of Electrical and Electronics Engineers Inc., pp. 393-397, 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2016, Heraklion, Crete, Greece, 2/17/16. https://doi.org/10.1109/PDP.2016.20
Berta A, Hegedus I, Jelasity M. Dimension Reduction Methods for Collaborative Mobile Gossip Learning. In Proceedings - 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2016. Institute of Electrical and Electronics Engineers Inc. 2016. p. 393-397. 7445364 https://doi.org/10.1109/PDP.2016.20
Berta, Arpad ; Hegedus, Istvan ; Jelasity, M. / Dimension Reduction Methods for Collaborative Mobile Gossip Learning. Proceedings - 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2016. Institute of Electrical and Electronics Engineers Inc., 2016. pp. 393-397
@inproceedings{604ce0bf891b4ac5949bdd63d36885d2,
title = "Dimension Reduction Methods for Collaborative Mobile Gossip Learning",
abstract = "Decentralized learning algorithms are very sensitive to the size of the raw data records due to the resulting large communication cost. This can, in the worst case, even make decentralized learning infeasible. Dimension reduction is a key technique to compress data and to obtain small models. In this paper, we propose a number of robust and efficient decentralized approaches to dimension reduction in the system model where each network node holds only one data record. These algorithms build on searching for good random projections. We present a thorough experimental comparison of the proposed algorithms and compare them with a variant of distributed singular value decomposition (SVD), a state-of-the-art algorithm for dimension reduction. We base our experiments on a trace of real mobile phone usage. We conclude that our method based on selecting good random projections is preferable and provides good quality results when the output is required on a very short timescale, within tens of minutes. We also present a hybrid method that combines the advantages of random projections and SVD. We demonstrate that the hybrid method offers good performance over all timescales.",
keywords = "dimension reduction, distributed data mining, gossip",
author = "Arpad Berta and Istvan Hegedus and M. Jelasity",
year = "2016",
month = "3",
day = "31",
doi = "10.1109/PDP.2016.20",
language = "English",
isbn = "9781467387750",
pages = "393--397",
booktitle = "Proceedings - 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2016",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Dimension Reduction Methods for Collaborative Mobile Gossip Learning

AU - Berta, Arpad

AU - Hegedus, Istvan

AU - Jelasity, M.

PY - 2016/3/31

Y1 - 2016/3/31

N2 - Decentralized learning algorithms are very sensitive to the size of the raw data records due to the resulting large communication cost. This can, in the worst case, even make decentralized learning infeasible. Dimension reduction is a key technique to compress data and to obtain small models. In this paper, we propose a number of robust and efficient decentralized approaches to dimension reduction in the system model where each network node holds only one data record. These algorithms build on searching for good random projections. We present a thorough experimental comparison of the proposed algorithms and compare them with a variant of distributed singular value decomposition (SVD), a state-of-the-art algorithm for dimension reduction. We base our experiments on a trace of real mobile phone usage. We conclude that our method based on selecting good random projections is preferable and provides good quality results when the output is required on a very short timescale, within tens of minutes. We also present a hybrid method that combines the advantages of random projections and SVD. We demonstrate that the hybrid method offers good performance over all timescales.

AB - Decentralized learning algorithms are very sensitive to the size of the raw data records due to the resulting large communication cost. This can, in the worst case, even make decentralized learning infeasible. Dimension reduction is a key technique to compress data and to obtain small models. In this paper, we propose a number of robust and efficient decentralized approaches to dimension reduction in the system model where each network node holds only one data record. These algorithms build on searching for good random projections. We present a thorough experimental comparison of the proposed algorithms and compare them with a variant of distributed singular value decomposition (SVD), a state-of-the-art algorithm for dimension reduction. We base our experiments on a trace of real mobile phone usage. We conclude that our method based on selecting good random projections is preferable and provides good quality results when the output is required on a very short timescale, within tens of minutes. We also present a hybrid method that combines the advantages of random projections and SVD. We demonstrate that the hybrid method offers good performance over all timescales.

KW - dimension reduction

KW - distributed data mining

KW - gossip

UR - http://www.scopus.com/inward/record.url?scp=84968914384&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84968914384&partnerID=8YFLogxK

U2 - 10.1109/PDP.2016.20

DO - 10.1109/PDP.2016.20

M3 - Conference contribution

AN - SCOPUS:84968914384

SN - 9781467387750

SP - 393

EP - 397

BT - Proceedings - 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing, PDP 2016

PB - Institute of Electrical and Electronics Engineers Inc.

ER -