Faster Algorithms for Max-Product Message-Passing

  title={Faster Algorithms for Max-Product Message-Passing},
  author={Julian J. McAuley and Tib{\'e}rio S. Caetano},
  journal={Journal of Machine Learning Research},
Maximum A Posteriorinference in graphical models is often solved via message-p assing algorithms, such as the junction-tree algorithm or loopy belief -propagation. The exact solution to this problem is well-known to be exponential in the size of the max i al cliques of the triangulated model, while approximate inference is typically exponenti al in the size… CONTINUE READING