GraphLab is an open source project to produce free implementations of scalable machine learning algorithms on multicore machine and clusters. The project was initiated in 2009 by Prof. Carlos Guestrin from Carnegie Mellon University and Prof. Joseph Hellerstein from University of California, Berkeley. GraphLab is a rapidly growing project with a large user base. As for October 2011, approximately 2000 institutions around the world installed Graphlab, including academia, research labs and industry.
As the amounts of collected data and computing power grows (multicore, GPUs, clusters, clouds), modern datasets no longer fit into one computing node. Efficient distributed/parallel algorithms for handling large scale data are required. The GraphLab framework is a parallel programming abstraction targeted for sparse iterative graph algorithms. GraphLab provides a high level programming interface, allowing a rapid deployment of distributed machine learning algorithms. The main design considerations behind GraphLab are: * Sparse data with local dependencies Iterative algorithms Potentially asynchronous execution
On top of GraphLab, several implemented libraries of algorithms: Collaborative filtering library - includes implementation of various matrix factorization techniques, including alternating least squares (ALS), weighted alternating least squares (wALS), non-negative matrix factorization (NMF), Bayesian probabilistic tensor factorization (BPTF), Probabilistic matrix factorization (PMF), Lanczos method, Singular Value Decomposition (SVD), Matrix decomposition with sparse factor matrices.
Clustering library - includes implementation of K-means, K-means++, Latent Dirichlet Allocation (LDA), K-Core decomposition, Item-KNN and user-KNN.
Linear solvers library - includes implementation of Jacobi method, Gaussian Belief Propagation, Shotgun LASSO, and Shotgun sparse logistic regression.
- Changes to previous version:
No one has posted any comments yet. Perhaps you'd like to be the first?
Leave a comment
You must be logged in to post comments.