The Community for Technology Leaders
Green Image
Issue No. 12 - Dec. (2012 vol. 34)
ISSN: 0162-8828
pp: 2365-2378
Jun Zhu , Dept. of Comput. Sci. & Technol., Tsinghua Univ., Beijing, China
Fuchun Sun , Dept. of Comput. Sci. & Technol., Tsinghua Univ., Beijing, China
E. P. Xing , Sch. of Comput. Sci., Carnegie Mellon Univ., Pittsburgh, PA, USA
Ning Chen , Dept. of Comput. Sci. & Technol., Tsinghua Univ., Beijing, China
Learning salient representations of multiview data is an essential step in many applications such as image classification, retrieval, and annotation. Standard predictive methods, such as support vector machines, often directly use all the features available without taking into consideration the presence of distinct views and the resultant view dependencies, coherence, and complementarity that offer key insights to the semantics of the data, and are therefore offering weak performance and are incapable of supporting view-level analysis. This paper presents a statistical method to learn a predictive subspace representation underlying multiple views, leveraging both multiview dependencies and availability of supervising side-information. Our approach is based on a multiview latent subspace Markov network (MN) which fulfills a weak conditional independence assumption that multiview observations and response variables are conditionally independent given a set of latent variables. To learn the latent subspace MN, we develop a large-margin approach which jointly maximizes data likelihood and minimizes a prediction loss on training data. Learning and inference are efficiently done with a contrastive divergence method. Finally, we extensively evaluate the large-margin latent MN on real image and hotel review datasets for classification, regression, image annotation, and retrieval. Our results demonstrate that the large-margin approach can achieve significant improvements in terms of prediction performance and discovering predictive latent subspace representations.
support vector machines, data analysis, image classification, image representation, image retrieval, learning (artificial intelligence), Markov processes, regression analysis, regression, large-margin predictive latent subspace learning, multiview data analysis, salient multiview data representations, image classification, image retrieval, image annotation, support vector machines, view-level analysis, statistical method, supervising side-information, multiview latent subspace Markov network, latent subspace MN, data likelihood maximization, contrastive divergence method, hotel review datasets, predictive latent subspace representations, Learning systems, Image retrieval, Classification, image retrieval and annotation, Latent subspace model, large-margin learning, classification, regression
Jun Zhu, Fuchun Sun, E. P. Xing, Ning Chen, "Large-Margin Predictive Latent Subspace Learning for Multiview Data Analysis", IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 34, no. , pp. 2365-2378, Dec. 2012, doi:10.1109/TPAMI.2012.64
99 ms
(Ver 3.3 (11022016))