Multimodal Music Search and Discovery

Professor Gert Lanckriet
UCSD

Friday, April 23, 2010 at 2:00pm
Engr IV Maxwell Room 57-124

Abstract
The revolution in production and distribution of music, which has made millions of audio clips instantly available to millions of people, has created the need for novel music search and discovery technologies. While successful technologies with great societal impact exist for text-based document search (e.g., Yahoo!, Google, etc.), a "Google for Music" has yet to stand up: there is no easy way to find a "mellow Beatles song" on a nostalgic night, "scary Halloween music" on October 31st, or address a sudden desire for "romantic jazz with saxophone and deep male vocals" without knowing an appropriate artist or song title.

The non-text-based, multimodal character of Internet-wide information about music (audio clips, lyrics, web documents, artist networks, band images, etc.) poses a new and difficult challenge to existing database technology, due to its dependence on unimodal, text-based data structures. Two fundamental research questions are at the core of addressing this challenge: 1) The automated indexing of non-text based music content and 2) the automated integration of the heterogeneous content of multimodal music databases, to retrieve the most relevant information, given a query.

In this talk, I will outline some of my recent research in machine learning, statistics and optimization, inspired and driven by the previous two research questions in the emerging field of computer audition and music information retrieval. This will cover a spectrum from sparse generalized eigenvalue problems to human computation games, and from clustering graphical models to multiple-kernel partial order embeddings.

Biography

Gert Lanckriet received a Master's degree in Electrical Engineering from the Katholieke Universiteit Leuven, Leuven, Belgium, in 2000 and the M.S. and Ph.D. degrees in Electrical Engineering and Computer Science from the University of California, Berkeley in 2001 respectively 2005. In 2005, he joined the Department of Electrical and Computer Engineering at the University of California, San Diego, where he heads the Computer Audition Laboratory. He was awarded the SIAM Optimization Prize in 2008 and is the recipient of a Hellman Fellowship in 2009. His research focuses on the interplay of convex optimization, machine learning and applied statistics, with applications in computer audition and music information retrieval.