My interests are in machine learning and its applications to multimedia, in
other words, using
statistical techniques to extract meaningful information from images,
video, text or sound. I'm also interested in multiple view learning: exploiting different
views of data in machine learning tasks, and convex optimization.
I'm part of the Computer Audition Lab
and the Computational Statistics and Machine
Learning group (COSMAL).
Finding words that are musically meaningful: We use a statistical
technique known as sparse canonical component analysis to find words
that are highly correlated to accompanying audio signals.
Modeling how humans describe music: We use Gaussian mixture models
to discover the audio structure behind songs that have been labeled
with keywords. A machine can use these models to annotate new songs
or to retrieve a list of songs based on a keyword query.
Douglas Turnbull, Luke Barrington, David Torres,
and Gert Lanckriet (2008).
Semantic Annotation and Retrieval of Music
and Sound Effects.
IEEE Transactions on Audio, Speech
and Language Processing, Volume: 16, Issue: 2.
David Torres, Bharath K. Sriperumbudur,
and Gert Lanckriet (2007).
Finding Musically MeaningfulWords
by Sparse CCA.
Neural Information Processing Systems (NIPS) Workshop on
Music, the Brain and Cognition.
Bharath K. Sriperumbudur, David Torres and Gert Lanckriet(2007).
Sparse Eigen Methods by D.C. Programming
International Conference on Machine Learning.
David Torres, Douglas Turnbull, Luke Barrington and Gert Lanckriet
Identifying Words that are Musically Meaningful
ISMIR, International Conference on Music Information Retrieval.
Douglas Turnbull, Luke Barrington,
David Torres and Gert Lanckriet (2007).
Towards Musical Query-by-Semantic Description
using the CAL500 Data Set
SIGIR, Special Interest Group on Information Retrieval.
You can download a pdf version of my resume here.