Bayesian Online Learning for Multi-label and Multi-variate Performance Measures

Xinhua Zhang, Thore Graepel, Ralf Herbrich
Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, PMLR 9:956-963, 2010.

Abstract

Many real world applications employ multi-variate performance measures and each example can belong to multiple classes. The currently most popular approaches train an SVM for each class, followed by ad hoc thresholding. Probabilistic models using Bayesian decision theory are also commonly adopted. In this paper, we propose a Bayesian online multi-label classification framework (BOMC) which learns a probabilistic linear classifier. The likelihood is modeled by a graphical model similar to TrueSkill$^\text{TM}$, and inference is based on Gaussian density filtering with expectation propagation. Using samples from the posterior, we label the testing data by maximizing the expected $F_1$-score. Our experiments on Reuters1-v2 dataset show BOMC compares favorably to the state-of-the-art online learners in macro-averaged $F_1$-score and training time.

Cite this Paper

BibTeX
@InProceedings{pmlr-v9-zhang10b, title = {Bayesian Online Learning for Multi-label and Multi-variate Performance Measures}, author = {Zhang, Xinhua and Graepel, Thore and Herbrich, Ralf}, booktitle = {Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics}, pages = {956--963}, year = {2010}, editor = {Teh, Yee Whye and Titterington, Mike}, volume = {9}, series = {Proceedings of Machine Learning Research}, address = {Chia Laguna Resort, Sardinia, Italy}, month = {13--15 May}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v9/zhang10b/zhang10b.pdf}, url = {https://proceedings.mlr.press/v9/zhang10b.html}, abstract = {Many real world applications employ multi-variate performance measures and each example can belong to multiple classes. The currently most popular approaches train an SVM for each class, followed by ad hoc thresholding. Probabilistic models using Bayesian decision theory are also commonly adopted. In this paper, we propose a Bayesian online multi-label classification framework (BOMC) which learns a probabilistic linear classifier. The likelihood is modeled by a graphical model similar to TrueSkill$^\text{TM}$, and inference is based on Gaussian density filtering with expectation propagation. Using samples from the posterior, we label the testing data by maximizing the expected $F_1$-score. Our experiments on Reuters1-v2 dataset show BOMC compares favorably to the state-of-the-art online learners in macro-averaged $F_1$-score and training time.} }
Endnote
%0 Conference Paper %T Bayesian Online Learning for Multi-label and Multi-variate Performance Measures %A Xinhua Zhang %A Thore Graepel %A Ralf Herbrich %B Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2010 %E Yee Whye Teh %E Mike Titterington %F pmlr-v9-zhang10b %I PMLR %P 956--963 %U https://proceedings.mlr.press/v9/zhang10b.html %V 9 %X Many real world applications employ multi-variate performance measures and each example can belong to multiple classes. The currently most popular approaches train an SVM for each class, followed by ad hoc thresholding. Probabilistic models using Bayesian decision theory are also commonly adopted. In this paper, we propose a Bayesian online multi-label classification framework (BOMC) which learns a probabilistic linear classifier. The likelihood is modeled by a graphical model similar to TrueSkill$^\text{TM}$, and inference is based on Gaussian density filtering with expectation propagation. Using samples from the posterior, we label the testing data by maximizing the expected $F_1$-score. Our experiments on Reuters1-v2 dataset show BOMC compares favorably to the state-of-the-art online learners in macro-averaged $F_1$-score and training time.
RIS
TY - CPAPER TI - Bayesian Online Learning for Multi-label and Multi-variate Performance Measures AU - Xinhua Zhang AU - Thore Graepel AU - Ralf Herbrich BT - Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics DA - 2010/03/31 ED - Yee Whye Teh ED - Mike Titterington ID - pmlr-v9-zhang10b PB - PMLR DP - Proceedings of Machine Learning Research VL - 9 SP - 956 EP - 963 L1 - http://proceedings.mlr.press/v9/zhang10b/zhang10b.pdf UR - https://proceedings.mlr.press/v9/zhang10b.html AB - Many real world applications employ multi-variate performance measures and each example can belong to multiple classes. The currently most popular approaches train an SVM for each class, followed by ad hoc thresholding. Probabilistic models using Bayesian decision theory are also commonly adopted. In this paper, we propose a Bayesian online multi-label classification framework (BOMC) which learns a probabilistic linear classifier. The likelihood is modeled by a graphical model similar to TrueSkill$^\text{TM}$, and inference is based on Gaussian density filtering with expectation propagation. Using samples from the posterior, we label the testing data by maximizing the expected $F_1$-score. Our experiments on Reuters1-v2 dataset show BOMC compares favorably to the state-of-the-art online learners in macro-averaged $F_1$-score and training time. ER -
APA
Zhang, X., Graepel, T. & Herbrich, R.. (2010). Bayesian Online Learning for Multi-label and Multi-variate Performance Measures. Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 9:956-963 Available from https://proceedings.mlr.press/v9/zhang10b.html.

Related Material