Past Awardee

Toward a Robust and Large-Scale Music Information Retrieval System: A Multidisciplinary Approach to Development and Evaluation

Stephen Downie
Stephen Downie

College: Graduate School of Library and Information Sciences
Award year: 2000-2001

The problems associated with the creation, deployment, and evaluation of robust, large-scale, and content-based Music Information Retrieval (MIR) systems are far from trivial. Music information is inherently multi-faceted, multi-representational (i.e., can be represented in many different ways), and space-intensive. The dizzyingly complex interaction of Pitch, Temporal, Harmonic, Timbral, Editorial, Textual, and Bibliographic facets, make music information difficult to store, and then retrieve, in any robust, large-scale, and comprehensive manner (i.e., this complex interaction is the "MIR problem"). The goal of this NCSA Fellows proposal is to draw upon the multi-disciplinary expertise of those at NCSA (and around UIUC) in dealing with other large-scale and complex information types. Through the implementation of a multi-disciplinary approach to MIR, I intend to identify, and then formally evaluate, the possible MIR application of those pre-existing supercomputing solutions currently being developed and implemented at NCSA.