IS = { zkontrolovano 25 Jan 2014 },
  UPDATE  = { 2013-08-16 },
  author =      {Sixta, Tom{\'a}{\v s} and Flach, Boris},
  title =       {Unsupervised (parameter) learning for MRFs on bipartite graphs},
  year =        {2013},
  pages =       {72.1-72.11},
  booktitle =   {Proceedings of the British Machine Vision Conference},
  publisher =   {BMVA},
  address =     {Imaging Science, Stopford Building, University of Manchester, Oxford, United Kingdom},
  editor =      {Burghardt, Tilo and Mayol-Cuevas, Walterio and Mirmehdi, Majid},
  book_pages =  {10},
  isbn =        {1-901725-49-9},
  month =       {September},
  day =         {9-13},
  venue =       {Bristol, United Kingdom},
  annote =      {We consider unsupervised (parameter) learning for
     general Markov random fields on bipartite graphs. This model
     class includes Restricted Boltzmann Machines. We show that
     besides the widely used stochastic gradient approximation
     (a.k.a. Persistent Con- trastive Divergence) there is an
     alternative learning approach - a modified EM algorithm which is
     tractable because of the bipartiteness of the model graph. We
     compare the re- sulting double loop algorithm and the PCD
     learning experimentally and show that the former converges faster
     and more stable than the latter.},
  keywords =    {MRF, parameter learning, pseudo-likelihood, EM algorithm},
  project =     {GACR P202/12/2071},