Yuan (Alan) Qi, Ph.D.
MIT Media Lab
20 Ames Street
I have moved to Purdue. > Shortly, you should automatically
be taken to my new homepage. If not, please follow the link
My research interests include approximate inference and
learning, model selection, nonparametric Bayesian methods,
nonlinear optimization, and
their applications in
functional genomics and wireless communications.
, Yuan Qi, Martin Szummer, and Thomas P. Minka,
To appear in AISTATS 2005. [paper/pdf
Predictive Automatic Relevance
Determination by Expectation Propagation, Yuan Qi, Thomas P. Minka,
Rosalind W. Picard, and Zoubin Ghahramani, in the Proceedings of
Conference on Machine Learning, July 4-8, 2004, Banff, Alberta, Canada.
sparse classifiers, which were applied to gene expression
by Expectation Propagation
, Thomas Minka and Yuan Qi, Neural
Processing Systems, December 2003, British Columbia, Canada. [pdf
efficient inference algorithm for loopy graphs.
Expectation Propagation for Signal Detection in Flat-fading Channels
Yuan Qi and Thomas Minka, MIT Media Lab Technical Report
Also, in the Proceedings of IEEE International Symposium on Information
Theory, June, 2003, Yokohama, Japan. [Abstract
fixed-lag smoothing algorithm for hybrid dynamic Bayesian networks with
its application to wireless communications.
Questions and answers about philosophy of
causation, and human/machine learning, Yuan Qi, October
Markov Chain Monte-Carlo Algorithms
, Yuan Qi and
Thomas P. Minka, First Cape Cod Workshop on Monte Carlo Methods,
2002, Cape Cod, Massachusetts. [slides/ps
optimization techniques with MCMC leads to new fast sampling methods
Context-sensitive Bayesian Classifiers and Application
to Mouse Pressure Pattern Classification, Yuan Qi, and
W. Picard, in Proceedings of International Conference on Pattern
August 2002, Québec City, Canada. [slide/ps]
probabilistic way to combine multiple classifiers which are trained on
different subsets of a given training set.
Bayesian Spectrum Estimation of Unevenly Sampled Nonstationary
Yuan Qi, Thomas P. Minka, and Rosalind W. Picard, MIT Media Lab
Report Vismod-TR-556, [Abstract]
Check out this web page
that summarizes experimental results, including comparison with
methods, e.g., Multitaper methods. The short version of this paper does
not include sparsification techniques and appears in ICASSP 02, May
Matlab implementation of our new spectrum estimation algorithm [download]
Hybrid Independent Component Analysis and Support Vector Machine
Learning Scheme for Face Detection, Y. Qi, D. DeMenthon, and D.
International Conference on Acoustics, Speech, and Signal Processing
May 2001, Salt Lake City,Utah. [ps]
Learning Algorithms for Video and Audio Processing: Independent
Analysis and Support Vector Machine based Approaches, Yuan
Technical Report LAMP-TR-056, CAR-TR-951, CS-TR-4174, Center for
Research, University of Maryland at College Park, August 2000.
Subband-based Independent Component Analysis, Yuan Qi, S.A.
P.S. Krishnaprasad, Proceedings of ICA2000,19-22 June 2000, Helsinki,
for graphical models, CMU CALD Machine learning lunch, April, 2004
I was a teaching assistant for MAS
622J Pattern Recognition in 2002. Besides my TA duty, I also did
lectures on Kalman filtering and smoothing, Junction tree algorithm,
Bayesian point machines.
Some photos I have taken in Spain (Valencia, Madrid, and
Japan (Kyoto and Tokyo), and US (Yellowstone).
E-mail: yuanqi (you can make the at sign) media (Dot) mit (doT) edu
Photo credit: Wei Chai
Last modified: Nov. 6, 2004