Learning Temporal Sequence Model from Partially Labeled Data. Shi, Y., Bobick, A., & Essa, I. In Computer Vision and Pattern Recognition, 2006 IEEE Computer Society Conference on, volume 2, pages 1631-1638, 2006.
doi  abstract   bibtex   
Graphical models are often used to represent and recognize activities. Purely unsupervised methods (such as HMMs) can be trained automatically but yield models whose internal structure - the nodes - are difficult to interpret semantically. Manually constructed networks typically have nodes corresponding to sub-events, but the programming and training of these networks is tedious and requires extensive domain expertise. In this paper, we propose a semi-supervised approach in which a manually structured, Propagation Network (a form of a DBN) is initialized from a small amount of fully annotated data, and then refined by an EM-based learning method in an unsupervised fashion. During node refinement (the M step) a boosting-based algorithm is employed to train the evidence detectors of individual nodes. Experiments on a variety of data types - vision and inertial measurements - in several tasks demonstrate the ability to learn from as little as one fully annotated example accompanied by a small number of positive but non-annotated training examples. The system is applied to both recognition and anomaly detection tasks.
@InProceedings{Shi2006,
  Title                    = {Learning Temporal Sequence Model from Partially Labeled Data},
  Author                   = {Shi, Yifan and Bobick, A. and Essa, I.},
  Booktitle                = {Computer Vision and Pattern Recognition, 2006 IEEE Computer Society Conference on},
  Year                     = {2006},
  Pages                    = {1631-1638},
  Volume                   = {2},

  Abstract                 = {Graphical models are often used to represent and recognize activities. Purely unsupervised methods (such as HMMs) can be trained automatically but yield models whose internal structure - the nodes - are difficult to interpret semantically. Manually constructed networks typically have nodes corresponding to sub-events, but the programming and training of these networks is tedious and requires extensive domain expertise. In this paper, we propose a semi-supervised approach in which a manually structured, Propagation Network (a form of a DBN) is initialized from a small amount of fully annotated data, and then refined by an EM-based learning method in an unsupervised fashion. During node refinement (the M step) a boosting-based algorithm is employed to train the evidence detectors of individual nodes. Experiments on a variety of data types - vision and inertial measurements - in several tasks demonstrate the ability to learn from as little as one fully annotated example accompanied by a small number of positive but non-annotated training examples. The system is applied to both recognition and anomaly detection tasks.},
  Doi                      = {10.1109/CVPR.2006.174},
  ISSN                     = {1063-6919},
  Keywords                 = {Application software;Boosting;Computer vision;Detectors;Graphical models;Hidden Markov models;Learning systems;Semisupervised learning;State-space methods;Surveillance},
  Timestamp                = {2014.12.22}
}
Downloads: 0