M3: Scaling Up Machine Learning via Memory Mapping
This addresses the challenge of handling large datasets for machine learning practitioners, offering a single-machine alternative to distributed systems, though it builds incrementally on prior virtual memory techniques for graph mining.
The paper tackles the problem of scaling machine learning algorithms to datasets that exceed RAM by proposing M3, a memory mapping approach that enables out-of-core processing, achieving speeds significantly faster than a 4-instance Spark cluster and comparable to an 8-instance cluster on datasets up to 190GB.
To process data that do not fit in RAM, conventional wisdom would suggest using distributed approaches. However, recent research has demonstrated virtual memory's strong potential in scaling up graph mining algorithms on a single machine. We propose to use a similar approach for general machine learning. We contribute: (1) our latest finding that memory mapping is also a feasible technique for scaling up general machine learning algorithms like logistic regression and k-means, when data fits in or exceeds RAM (we tested datasets up to 190GB); (2) an approach, called M3, that enables existing machine learning algorithms to work with out-of-core datasets through memory mapping, achieving a speed that is significantly faster than a 4-instance Spark cluster, and comparable to an 8-instance cluster.