NCSA Music Mining Project among HPC Analytics Finalists
released 11.11.05
Contact
Trish Barker
NCSA Public Information Specialist
tlbarker@ncsa.uiuc.edu
217.265.8013
URBANA, IL —
A project that uses data mining technology developed at NCSA is among the six finalists in the HPC Analytics Challenge at SC05 (Nov. 12-18 in Seattle). The new challenge is designed to honor top technical and commercial applications developed from leading-edge, advanced analytics techniques to solve complex, real-world problems.
The terascale music mining project aims to create secure, accessible terascale collections of music materials in a variety of audio, symbolic, and metadata forms. Researchers from NCSA and the Graduate School of Library and Information Science at the University of Illinois at Urbana-Champaign developed a music-specific extension to its D2K (Data to Knowledge) machine learning and data mining environment. The extension, dubbed M2K, comprises approximately 200 music-specific feature extraction and music mining modules. More information on M2K is available at http://music-ir.org/evaluation/m2k/.
At SC05, two M2K applications will be presented: real-time genre classification and audio artist identification. In both cases, the M2K system assesses music audio samples, identifies them, and continues to learn and refine its judgments. A video demonstration, featuring the real-time genre classification, is available at http://music-ir.org/m2kvid/.
The six HPC Analytics finalists will present their applications from 1:30 to 5 p.m. on Tuesday, Nov. 15. The awards ceremony will be held at 3:30 p.m. on Thursday, Nov. 17. For more information, go to http://sc05.supercomputing.org/initiatives/home.php.
| 
