GraphLab: A New Framework for Parallel Machine Learning

Yucheng Low, Joseph Gonzalez, Aapo Kyrola, Danny Bickson, Carlos Guestrin
Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence, PMLR R8:347-356, 2010.

Abstract

Designing and implementing efficient, provably correct parallel machine learning (ML) algo- rithms is challenging. Existing high-level par- allel abstractions like MapReduce are insuf- ficiently expressive while low-level tools like MPI and Pthreads leave ML experts repeatedly solving the same design challenges. By tar- geting common patterns in ML, we developed GraphLab, which improves upon abstractions like MapReduce by compactly expressing asyn- chronous iterative algorithms with sparse com- putational dependencies while ensuring data con- sistency and achieving a high degree of parallel performance. We demonstrate the expressiveness of the GraphLab framework by designing and implementing parallel versions of belief propaga- tion, Gibbs sampling, Co-EM, Lasso and Com- pressed Sensing. We show that using GraphLab we can achieve excellent parallel performance on large scale real-world problems.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR8-low10a, title = {GraphLab: A New Framework for Parallel Machine Learning}, author = {Low, Yucheng and Gonzalez, Joseph and Kyrola, Aapo and Bickson, Danny and Guestrin, Carlos}, booktitle = {Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence}, pages = {347--356}, year = {2010}, editor = {Grünwald, Peter and Spirtes, Peter}, volume = {R8}, series = {Proceedings of Machine Learning Research}, month = {08--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r8/main/assets/low10a/low10a.pdf}, url = {https://proceedings.mlr.press/r8/low10a.html}, abstract = {Designing and implementing efficient, provably correct parallel machine learning (ML) algo- rithms is challenging. Existing high-level par- allel abstractions like MapReduce are insuf- ficiently expressive while low-level tools like MPI and Pthreads leave ML experts repeatedly solving the same design challenges. By tar- geting common patterns in ML, we developed GraphLab, which improves upon abstractions like MapReduce by compactly expressing asyn- chronous iterative algorithms with sparse com- putational dependencies while ensuring data con- sistency and achieving a high degree of parallel performance. We demonstrate the expressiveness of the GraphLab framework by designing and implementing parallel versions of belief propaga- tion, Gibbs sampling, Co-EM, Lasso and Com- pressed Sensing. We show that using GraphLab we can achieve excellent parallel performance on large scale real-world problems.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T GraphLab: A New Framework for Parallel Machine Learning %A Yucheng Low %A Joseph Gonzalez %A Aapo Kyrola %A Danny Bickson %A Carlos Guestrin %B Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2010 %E Peter Grünwald %E Peter Spirtes %F pmlr-vR8-low10a %I PMLR %P 347--356 %U https://proceedings.mlr.press/r8/low10a.html %V R8 %X Designing and implementing efficient, provably correct parallel machine learning (ML) algo- rithms is challenging. Existing high-level par- allel abstractions like MapReduce are insuf- ficiently expressive while low-level tools like MPI and Pthreads leave ML experts repeatedly solving the same design challenges. By tar- geting common patterns in ML, we developed GraphLab, which improves upon abstractions like MapReduce by compactly expressing asyn- chronous iterative algorithms with sparse com- putational dependencies while ensuring data con- sistency and achieving a high degree of parallel performance. We demonstrate the expressiveness of the GraphLab framework by designing and implementing parallel versions of belief propaga- tion, Gibbs sampling, Co-EM, Lasso and Com- pressed Sensing. We show that using GraphLab we can achieve excellent parallel performance on large scale real-world problems. %Z Reissued by PMLR on 04 October 2026.
APA
Low, Y., Gonzalez, J., Kyrola, A., Bickson, D. & Guestrin, C.. (2010). GraphLab: A New Framework for Parallel Machine Learning. Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R8:347-356 Available from https://proceedings.mlr.press/r8/low10a.html. Reissued by PMLR on 04 October 2026.

Related Material