Dr.Srinivasan Parthasarathy has received the highly competitive IBM Faculty Award for 2007 for his work on Architecture Conscious Data Analysis and Management. Although the challenge of efficient use of hardware has been a major concern in the fields of database management and data mining for many years, it is only recently that significant efforts in the community have been spent on this problem. Careful algorithmic restructuring coupled with sound methods to explicitly leverage architectural features are essential to enable one to realize performance that is commensurate with emerging hardware technology.<\p>
Over the last couple of years, Professor Parthasarathy and his students have been exploring this problem domain and have successfully deployed architecture conscious solutions for key data mining algorithms such as association rule mining, tree mining and graph mining as well as in the context of indexing XML data. Articles describing this work have appeared at top conferences such as VLDB, SIGKDD, ICDM and CIKM and have been well received. In fact two of these articles were nominated for awards at these conferences (VLDB 2005, SIGKDD 2006) and one received the best research paper award. More details on this work can be found at the Data Mining Research Lab's website
As part of the IBM faculty award Professor Parthasarathy will be investigating the performance of key data mining and indexing kernels on the Sony-Toshiba-IBM Cell Broadband Engine Architecture. The Cell architecture is an exciting new high performance architecture that grew from the challenge to provide a cost and power effective solution for the gaming industry (it is the processor within the Sony PlayStation console). However, as recent research has shown the Cell architecture promises to have a much broader impact than what was originally intended, i.e., beyond the gaming industry.<\p>
The IBM Faculty awards program is a competitive international program intended to foster collaboration between researchers at leading universities worldwide and those at IBM research. To qualify for this program candidates must have an outstanding reputation for contributions in their field and show unusual promise.<\p>