Mercurial > cortex

comparison MIT-media-projects.org @ 334:c264ebf683b4

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
cleanup.
author Robert McIntyre <rlm@mit.edu>
date Fri, 20 Jul 2012 11:22:21 -0500
parents f4ef73370da1
children
comparison
equal deleted inserted replaced
333:f4ef73370da1 334:c264ebf683b4
1 *Machine Learning and Pattern Recognition with Multiple Modalities 1 *Machine Learning and Pattern Recognition with Multiple
2 Hyungil Ahn and Rosalind W. Picard 2 Modalities Hyungil Ahn and Rosalind W. Picard
3 3
4 This project develops new theory and algorithms to enable computers to 4 This project develops new theory and algorithms to enable
5 make rapid and accurate inferences from multiple modes of data, such 5 computers to make rapid and accurate inferences from
6 as determining a person's affective state from multiple sensors—video, 6 multiple modes of data, such as determining a person's
7 mouse behavior, chair pressure patterns, typed selections, or 7 affective state from multiple sensors--video, mouse behavior,
8 physiology. Recent efforts focus on understanding the level of a 8 chair pressure patterns, typed selections, or
9 person's attention, useful for things such as determining when to 9 physiology. Recent efforts focus on understanding the level
10 interrupt. Our approach is Bayesian: formulating probabilistic models 10 of a person's attention, useful for things such as
11 on the basis of domain knowledge and training data, and then 11 determining when to interrupt. Our approach is Bayesian:
12 performing inference according to the rules of probability 12 formulating probabilistic models on the basis of domain
13 theory. This type of sensor fusion work is especially challenging due 13 knowledge and training data, and then performing inference
14 to problems of sensor channel drop-out, different kinds of noise in 14 according to the rules of probability theory. This type of
15 different channels, dependence between channels, scarce and sometimes 15 sensor fusion work is especially challenging due to problems
16 inaccurate labels, and patterns to detect that are inherently 16 of sensor channel drop-out, different kinds of noise in
17 time-varying. We have constructed a variety of new algorithms for 17 different channels, dependence between channels, scarce and
18 solving these problems and demonstrated their performance gains over 18 sometimes inaccurate labels, and patterns to detect that are
19 other state-of-the-art methods. 19 inherently time-varying. We have constructed a variety of
20 new algorithms for solving these problems and demonstrated
21 their performance gains over other state-of-the-art methods.
20 22
21 http://affect.media.mit.edu/projectpages/multimodal/ 23 http://affect.media.mit.edu/projectpages/multimodal/
24