Dance Music Classification Using Inner Metric Analysis

Abstract

ALPS is a framework for implementing and parallelizing tree searchalgorithms. It employs a number of features to improve scalabilityand is designed specifically to support the implementation of dataintensive algorithms, in which large amounts of knowledge are generatedand must be maintained and shared during the search. Implementingsuch algorithms in a scalable manner is challenging both becauseof storage requirements and because of communications overhead incurredin the sharing of data. In this abstract, we describe the designof ALPS and how the design addresses these challenges. We presenttwo sample applications built with ALPS and preliminary computationalresults.