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 paper proposes projective dictionary pair learning (DPL) to jointly learn a synthesis and analysis dictionary so as to apply the dictionary learning approach to the classification setting while also enabling more efficient computation. Overall the paper proposes an interesting method. Even though the proposed method is a straight-forward extension of DL, the empirical results (esp the computational gains) looks quite promising. It would have been great to see: i) an accompanying theoretical analysis esp around convergence ii) analysis of the complexity of method as sample size increases (and hence with increasing m) It is interesting how the method imposes sparsity indirectly (as seen in Sec 2.4) by training $p^*_k$ to produce small coefficients.