Linear Time Solver for Primal SVM

Feiping Nie, Yizhen Huang, Heng Huang
Proceedings of the 31st International Conference on Machine Learning, PMLR 32(2):505-513, 2014.

Abstract

Support Vector Machines (SVM) is among the most popular classification techniques in machine learning, hence designing fast primal SVM algorithms for large-scale datasets is a hot topic in recent years. This paper presents a new L2-norm regularized primal SVM solver using Augmented Lagrange Multipliers, with linear-time computational cost for Lp-norm loss functions. The most computationally intensive steps (that determine the algorithmic complexity) of the proposed algorithm is purely and simply matrix-by-vector multiplication, which can be easily parallelized on a multi-core server for parallel computing. We implement and integrate our algorithm into the interfaces and framework of the well-known LibLinear software toolbox. Experiments show that our algorithm is with stable performance and on average faster than the state-of-the-art solvers such as SVMperf , Pegasos and the LibLinear that integrates the TRON, PCD and DCD algorithms.

Cite this Paper


BibTeX
@InProceedings{pmlr-v32-niea14, title = {Linear Time Solver for Primal SVM}, author = {Nie, Feiping and Huang, Yizhen and Huang, Heng}, booktitle = {Proceedings of the 31st International Conference on Machine Learning}, pages = {505--513}, year = {2014}, editor = {Xing, Eric P. and Jebara, Tony}, volume = {32}, number = {2}, series = {Proceedings of Machine Learning Research}, address = {Bejing, China}, month = {22--24 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v32/niea14.pdf}, url = {https://proceedings.mlr.press/v32/niea14.html}, abstract = {Support Vector Machines (SVM) is among the most popular classification techniques in machine learning, hence designing fast primal SVM algorithms for large-scale datasets is a hot topic in recent years. This paper presents a new L2-norm regularized primal SVM solver using Augmented Lagrange Multipliers, with linear-time computational cost for Lp-norm loss functions. The most computationally intensive steps (that determine the algorithmic complexity) of the proposed algorithm is purely and simply matrix-by-vector multiplication, which can be easily parallelized on a multi-core server for parallel computing. We implement and integrate our algorithm into the interfaces and framework of the well-known LibLinear software toolbox. Experiments show that our algorithm is with stable performance and on average faster than the state-of-the-art solvers such as SVMperf , Pegasos and the LibLinear that integrates the TRON, PCD and DCD algorithms.} }
Endnote
%0 Conference Paper %T Linear Time Solver for Primal SVM %A Feiping Nie %A Yizhen Huang %A Heng Huang %B Proceedings of the 31st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2014 %E Eric P. Xing %E Tony Jebara %F pmlr-v32-niea14 %I PMLR %P 505--513 %U https://proceedings.mlr.press/v32/niea14.html %V 32 %N 2 %X Support Vector Machines (SVM) is among the most popular classification techniques in machine learning, hence designing fast primal SVM algorithms for large-scale datasets is a hot topic in recent years. This paper presents a new L2-norm regularized primal SVM solver using Augmented Lagrange Multipliers, with linear-time computational cost for Lp-norm loss functions. The most computationally intensive steps (that determine the algorithmic complexity) of the proposed algorithm is purely and simply matrix-by-vector multiplication, which can be easily parallelized on a multi-core server for parallel computing. We implement and integrate our algorithm into the interfaces and framework of the well-known LibLinear software toolbox. Experiments show that our algorithm is with stable performance and on average faster than the state-of-the-art solvers such as SVMperf , Pegasos and the LibLinear that integrates the TRON, PCD and DCD algorithms.
RIS
TY - CPAPER TI - Linear Time Solver for Primal SVM AU - Feiping Nie AU - Yizhen Huang AU - Heng Huang BT - Proceedings of the 31st International Conference on Machine Learning DA - 2014/06/18 ED - Eric P. Xing ED - Tony Jebara ID - pmlr-v32-niea14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 32 IS - 2 SP - 505 EP - 513 L1 - http://proceedings.mlr.press/v32/niea14.pdf UR - https://proceedings.mlr.press/v32/niea14.html AB - Support Vector Machines (SVM) is among the most popular classification techniques in machine learning, hence designing fast primal SVM algorithms for large-scale datasets is a hot topic in recent years. This paper presents a new L2-norm regularized primal SVM solver using Augmented Lagrange Multipliers, with linear-time computational cost for Lp-norm loss functions. The most computationally intensive steps (that determine the algorithmic complexity) of the proposed algorithm is purely and simply matrix-by-vector multiplication, which can be easily parallelized on a multi-core server for parallel computing. We implement and integrate our algorithm into the interfaces and framework of the well-known LibLinear software toolbox. Experiments show that our algorithm is with stable performance and on average faster than the state-of-the-art solvers such as SVMperf , Pegasos and the LibLinear that integrates the TRON, PCD and DCD algorithms. ER -
APA
Nie, F., Huang, Y. & Huang, H.. (2014). Linear Time Solver for Primal SVM. Proceedings of the 31st International Conference on Machine Learning, in Proceedings of Machine Learning Research 32(2):505-513 Available from https://proceedings.mlr.press/v32/niea14.html.

Related Material