We aim to build a library of Machine Learning algorithms and a toolkit for writing new ones. Our toolkit is implemented using the DryadLINQ distributed computation framework. This enables the algorithms to process very large amounts of data using computer clusters.

The Large Vector library provides a very simple API, similar to map-reduce, for processing large distributed collections of numeric data.