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. This paper introduces belief propagation recursive neural networks. This kind of networks extends the feedforward (bottom-up) recursive networks with a backward (top-down) step during inference. This allows phrase level predictions and embeddings to give feedback to word embeddings and labels. This paper is overall well written.