Skip to main content

The matrix stick-breaking process for flexible multi-task learning

Publication ,  Journal Article
Xue, Y; Dunson, D; Carin, L
Published in: ACM International Conference Proceeding Series
August 23, 2007

In multi-task learning our goal is to design regression or classification models for each of the tasks and appropriately share information between tasks. A Dirichlet process (DP) prior can be used to encourage task clustering. However, the DP prior does not allow local clustering of tasks with respect to a subset of the feature vector without making independence assumptions. Motivated by this problem, we develop a new multitask-learning prior, termed the matrix stick-breaking process (MSBP), which encourages cross-task sharing of data. However, the MSBP allows separate clustering and borrowing of information for the different feature components. This is important when tasks are more closely related for certain features than for others. Bayesian inference proceeds by a Gibbs sampling algorithm and the approach is illustrated using a simulated example and a multi-national application.

Duke Scholars

Published In

ACM International Conference Proceeding Series

DOI

Publication Date

August 23, 2007

Volume

227

Start / End Page

1063 / 1070
 

Citation

APA
Chicago
ICMJE
MLA
NLM
Xue, Y., Dunson, D., & Carin, L. (2007). The matrix stick-breaking process for flexible multi-task learning. ACM International Conference Proceeding Series, 227, 1063–1070. https://doi.org/10.1145/1273496.1273630
Xue, Y., D. Dunson, and L. Carin. “The matrix stick-breaking process for flexible multi-task learning.” ACM International Conference Proceeding Series 227 (August 23, 2007): 1063–70. https://doi.org/10.1145/1273496.1273630.
Xue Y, Dunson D, Carin L. The matrix stick-breaking process for flexible multi-task learning. ACM International Conference Proceeding Series. 2007 Aug 23;227:1063–70.
Xue, Y., et al. “The matrix stick-breaking process for flexible multi-task learning.” ACM International Conference Proceeding Series, vol. 227, Aug. 2007, pp. 1063–70. Scopus, doi:10.1145/1273496.1273630.
Xue Y, Dunson D, Carin L. The matrix stick-breaking process for flexible multi-task learning. ACM International Conference Proceeding Series. 2007 Aug 23;227:1063–1070.

Published In

ACM International Conference Proceeding Series

DOI

Publication Date

August 23, 2007

Volume

227

Start / End Page

1063 / 1070