Corpus ID: 12918282

Concavity of reweighted Kikuchi approximation

@inproceedings{Loh2014ConcavityOR,
  title={Concavity of reweighted Kikuchi approximation},
  author={Po-Ling Loh and Andre Wibisono},
  booktitle={NIPS},
  year={2014}
}
We analyze a reweighted version of the Kikuchi approximation for estimating the log partition function of a product distribution defined over a region graph. We establish sufficient conditions for the concavity of our reweighted objective function in terms of weight assignments in the Kikuchi expansion, and show that a reweighted version of the sum product algorithm applied to the Kikuchi region graph will produce global optima of the Kikuchi approximation whenever the algorithm converges. When… Expand
1 Citations

References

SHOWING 1-10 OF 37 REFERENCES
Estimation and Marginalization Using the Kikuchi Approximation Methods
  • 52
  • Highly Influential
  • PDF
Convergence Analysis of Reweighted Sum-Product Algorithms
  • 47
  • PDF
Convergence Analysis of Reweighted Sum-Product Algorithms
  • 8
  • PDF
The Bethe Partition Function of Log-supermodular Graphical Models
  • N. Ruozzi
  • Computer Science, Mathematics
  • NIPS
  • 2012
  • 60
  • PDF
The Bethe Permanent of a Nonnegative Matrix
  • P. Vontobel
  • Mathematics, Computer Science
  • IEEE Transactions on Information Theory
  • 2013
  • 67
A new class of upper bounds on the log partition function
  • 446
  • Highly Influential
  • PDF
Loop Series and Bethe Variational Bounds in Attractive Graphical Models
  • 68
  • PDF
The Generalized Distributive Law and Free Energy Minimization
  • 81
  • PDF
Sufficient Conditions for Convergence of the Sum–Product Algorithm
  • J. Mooij, H. Kappen
  • Mathematics, Computer Science
  • IEEE Transactions on Information Theory
  • 2007
  • 228
  • PDF
Primal View on Belief Propagation
  • 9
  • PDF
...
1
2
3
4
...