File(s) under permanent embargo
Learning discriminative sequence models from partially labelled data for activity recognition
conference contribution
posted on 2008-01-01, 00:00 authored by T Truyen, H Bui, Quoc-Dinh Phung, Svetha VenkateshSvetha VenkateshRecognising daily activity patterns of people from low-level sensory data is an important problem. Traditional approaches typically rely on generative models such as the hidden Markov models and training on fully labelled data. While activity data can be readily acquired from pervasive sensors, e.g. in smart environments, providing manual labels to support fully supervised learning is often expensive. In this paper, we propose a new approach based on partially-supervised training of discriminative sequence models such as the conditional random field (CRF) and the maximum entropy Markov model (MEMM). We show that the approach can reduce labelling effort, and at the same time, provides us with the flexibility and accuracy of the discriminative framework. Our experimental results in the video surveillance domain illustrate that these models can perform better than their generative counterpart (i.e. the partially hidden Markov model), even when a substantial amount of labels are unavailable.
History
Event
Pacific Rim International Conference on Artificial Intelligence (10th : 2008 : Hanoi, Vietnam)Source
PRICAI 2008 : trends in artificial intelligence : 10th Pacific Rim International Conference on Artificial Intelligence, Hanoi, Vietnam, December 15-19, 2008, proceedingsSeries
Lecture notes in artificial intelligence ; 5351Pagination
903 - 912Publisher
SpringerLocation
Hanoi, VietnamPlace of publication
Berlin, GermanyPublisher DOI
Start date
2008-12-15End date
2008-12-19ISSN
0302-9743ISBN-13
9783540891963ISBN-10
354089196XLanguage
engPublication classification
E1.1 Full written paper - refereedCopyright notice
2008, SpringerExtent
117Editor/Contributor(s)
T Ho, Z ZhouTitle of proceedings
PRICAI 2008 : Trends in Artificial Intelligence : 10th Pacific Rim International Conference on Artificial Intelligence, Hanoi, Vietnam, December 15-19, 2008, proceedingsUsage metrics
Categories
No categories selectedKeywords
Licence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC