PROJECT TITLE :
Scalable Distributed Nonnegative Matrix Factorization with Block-Wise Updates - 2018
Nonnegative Matrix Factorization (NMF) has been applied with nice success on a big selection of applications. As NMF is increasingly applied to massive datasets like internet-scale dyadic information, it's desirable to leverage a cluster of machines to store those datasets and to hurry up the factorization method. But, it's difficult to efficiently implement NMF in an exceedingly distributed setting. During this Project, we show that by leveraging a replacement form of update functions, we can perform local aggregation and fully explore parallelism. Thus, the new kind is much additional efficient than the traditional type in distributed implementations. Moreover, below the new type of update functions, we will perform frequent updates and lazy updates, which aim to use the foremost recently updated data whenever possible and avoid unnecessary computations. Thus, frequent updates and lazy updates are a lot of efficient than their ancient concurrent counterparts. Through a series of experiments on a local cluster also because the Amazon EC2 cloud, we tend to demonstrate that our implementations with frequent updates or lazy updates are up to two orders of magnitude faster than the prevailing implementation with the ancient type of update functions.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here