Export Reviews, Discussions, Author Feedback and Meta-Reviews

First provide a summary of the paper, and then address the following criteria: Quality, clarity, originality and significance. The authors consider the problem of text classification using methods beyond standard bag-of-word approaches. The basic technique is to use an embedding of a document that incorporates latent information and produces a sum of kernels representation. Buidling on top of the representation for documents, classification is performed with a support measure machine. The combination of the representation and the SMM is applied to standard corpora and compared to other relevant approaches.