rlm@333: *Machine Learning and Pattern Recognition with Multiple Modalities
rlm@333: Hyungil Ahn and Rosalind W. Picard
rlm@333: 
rlm@333: This project develops new theory and algorithms to enable computers to
rlm@333: make rapid and accurate inferences from multiple modes of data, such
rlm@333: as determining a person's affective state from multiple sensors—video,
rlm@333: mouse behavior, chair pressure patterns, typed selections, or
rlm@333: physiology. Recent efforts focus on understanding the level of a
rlm@333: person's attention, useful for things such as determining when to
rlm@333: interrupt. Our approach is Bayesian: formulating probabilistic models
rlm@333: on the basis of domain knowledge and training data, and then
rlm@333: performing inference according to the rules of probability
rlm@333: theory. This type of sensor fusion work is especially challenging due
rlm@333: to problems of sensor channel drop-out, different kinds of noise in
rlm@333: different channels, dependence between channels, scarce and sometimes
rlm@333: inaccurate labels, and patterns to detect that are inherently
rlm@333: time-varying. We have constructed a variety of new algorithms for
rlm@333: solving these problems and demonstrated their performance gains over
rlm@333: other state-of-the-art methods.
rlm@333: 
rlm@333: http://affect.media.mit.edu/projectpages/multimodal/