Timezone: »
We propose a Bayesian undirected graphical model for semi-supervised multi-view learning (BMVL). This model makes explicit the previously unstated assumptions of a large class of semi-supervised multi-view learning algorithms, and clarifies the circumstances under which these assumptions fail. Building upon new insights from this model, we propose an improved method for BMVL: in particular, we derive a novel co-training kernel for Gaussian Process classifiers. Unlike some previous methods for multi-view learning, the resulting approach is convex and avoids local-maxima problems. Further, it automatically estimates how much each view should be trusted, and thus accommodates noisy or unreliable views. Experiments show that the approach is more accurate than previous multi-view learning algorithms.
Author Information
Shipeng Yu (University of Munich)
Balaji R Krishnapuram (IBM)
Romer E Rosales (University of Toronto)
Harald Steck (ETH Zurich)
R. Bharat Rao
Related Events (a corresponding poster, oral, or spotlight)
-
2007 Poster: Bayesian Multi-View Learning »
Wed. Dec 5th 06:30 -- 06:40 PM Room
More from the Same Authors
-
2013 Workshop: Machine Learning for Clinical Data Analysis and Healthcare »
Jenna Wiens · Finale P Doshi-Velez · Can Ye · Madalina Fiterau · Shipeng Yu · Le Lu · Balaji R Krishnapuram -
2011 Workshop: From statistical genetics to predictive models in personalized medicine »
Karsten Borgwardt · Oliver Stegle · Shipeng Yu · Glenn Fung · Faisal Farooq · Balaji R Krishnapuram -
2010 Workshop: Predictive Models in Personalized Medicine »
Faisal Farooq · Glenn Fung · Romer Rosales · Shipeng Yu · Jude W Shavlik · Balaji R Krishnapuram · Raju Kucherlapati -
2008 Workshop: Cost Sensitive Learning »
Balaji R Krishnapuram · Shipeng Yu · Oksana Yakhnenko · R. Bharat Rao · Lawrence Carin -
2007 Poster: On Ranking in Survival Analysis: Bounds on the Concordance Index »
Vikas C Raykar · Harald Steck · Balaji R Krishnapuram · Cary Dehing-Oberije · Philippe Lambin -
2006 Demonstration: Computer Aided Diagnosis of Early Stage Cancer »
Marcos Salganicoff · Luca Bogoni · R. Bharat Rao · Murat Dundar · Balaji R Krishnapuram · Glenn Fung -
2006 Poster: Multiple Instance Learning for Computer Aided Diagnosis »
Glenn Fung · Murat Dundar · Balaji R Krishnapuram · R. Bharat Rao -
2006 Poster: Gaussian Process Models for Discriminative Link Prediction »
Kai Yu · Wei Chu · Shipeng Yu · Volker Tresp · Zhao Xu