Transductive Learning of Structural SVMs via Prior Knowledge Constraints

Chun-Nam Yu
; Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics, PMLR 22:1367-1376, 2012.

Abstract

Reducing the number of labeled examples required to learn accurate prediction models is an important problem in structured output prediction. In this paper we propose a new transductive structural SVM algorithm that learns by incorporating prior knowledge constraints on unlabeled data. Our formulation supports different types of prior knowledge constraints, and can be trained efficiently. Experiments on two citation and advertisement segmentation tasks show that our transductive structural SVM can learn effectively from unlabeled data, achieving similar prediction accuracies when compared against other state-of-art algorithms.

Cite this Paper


BibTeX
@InProceedings{pmlr-v22-yu12, title = {Transductive Learning of Structural SVMs via Prior Knowledge Constraints}, author = {Chun-Nam Yu}, booktitle = {Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics}, pages = {1367--1376}, year = {2012}, editor = {Neil D. Lawrence and Mark Girolami}, volume = {22}, series = {Proceedings of Machine Learning Research}, address = {La Palma, Canary Islands}, month = {21--23 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v22/yu12/yu12.pdf}, url = {http://proceedings.mlr.press/v22/yu12.html}, abstract = {Reducing the number of labeled examples required to learn accurate prediction models is an important problem in structured output prediction. In this paper we propose a new transductive structural SVM algorithm that learns by incorporating prior knowledge constraints on unlabeled data. Our formulation supports different types of prior knowledge constraints, and can be trained efficiently. Experiments on two citation and advertisement segmentation tasks show that our transductive structural SVM can learn effectively from unlabeled data, achieving similar prediction accuracies when compared against other state-of-art algorithms.} }
Endnote
%0 Conference Paper %T Transductive Learning of Structural SVMs via Prior Knowledge Constraints %A Chun-Nam Yu %B Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2012 %E Neil D. Lawrence %E Mark Girolami %F pmlr-v22-yu12 %I PMLR %J Proceedings of Machine Learning Research %P 1367--1376 %U http://proceedings.mlr.press %V 22 %W PMLR %X Reducing the number of labeled examples required to learn accurate prediction models is an important problem in structured output prediction. In this paper we propose a new transductive structural SVM algorithm that learns by incorporating prior knowledge constraints on unlabeled data. Our formulation supports different types of prior knowledge constraints, and can be trained efficiently. Experiments on two citation and advertisement segmentation tasks show that our transductive structural SVM can learn effectively from unlabeled data, achieving similar prediction accuracies when compared against other state-of-art algorithms.
RIS
TY - CPAPER TI - Transductive Learning of Structural SVMs via Prior Knowledge Constraints AU - Chun-Nam Yu BT - Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics PY - 2012/03/21 DA - 2012/03/21 ED - Neil D. Lawrence ED - Mark Girolami ID - pmlr-v22-yu12 PB - PMLR SP - 1367 DP - PMLR EP - 1376 L1 - http://proceedings.mlr.press/v22/yu12/yu12.pdf UR - http://proceedings.mlr.press/v22/yu12.html AB - Reducing the number of labeled examples required to learn accurate prediction models is an important problem in structured output prediction. In this paper we propose a new transductive structural SVM algorithm that learns by incorporating prior knowledge constraints on unlabeled data. Our formulation supports different types of prior knowledge constraints, and can be trained efficiently. Experiments on two citation and advertisement segmentation tasks show that our transductive structural SVM can learn effectively from unlabeled data, achieving similar prediction accuracies when compared against other state-of-art algorithms. ER -
APA
Yu, C.. (2012). Transductive Learning of Structural SVMs via Prior Knowledge Constraints. Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics, in PMLR 22:1367-1376

Related Material