GraphLab: A New Framework for Parallel Machine Learning

TL;DR

GraphLab is a novel parallel framework tailored for machine learning algorithms, enabling efficient, correct, and expressive asynchronous computations with sparse dependencies, demonstrated on various ML tasks.

Contribution

Introduces GraphLab, a high-level parallel abstraction specifically designed for ML algorithms, improving expressiveness and performance over existing tools.

Findings

Successfully implemented parallel belief propagation, Gibbs sampling, Co-EM, Lasso, and Compressed Sensing.

Achieved high parallel performance on large-scale real-world problems.

Demonstrated expressiveness and efficiency of GraphLab for diverse ML algorithms.

Abstract

Designing and implementing efficient, provably correct parallel machine learning (ML) algorithms is challenging. Existing high-level parallel abstractions like MapReduce are insufficiently expressive while low-level tools like MPI and Pthreads leave ML experts repeatedly solving the same design challenges. By targeting common patterns in ML, we developed GraphLab, which improves upon abstractions like MapReduce by compactly expressing asynchronous iterative algorithms with sparse computational dependencies while ensuring data consistency and achieving a high degree of parallel performance. We demonstrate the expressiveness of the GraphLab framework by designing and implementing parallel versions of belief propagation, Gibbs sampling, Co-EM, Lasso and Compressed Sensing. We show that using GraphLab we can achieve excellent parallel performance on large scale real-world problems.