An Alternative View of Survey Propagation for Satisfiability
Phase Transitions in Computation and Reconstruction
March 09, 2005 11:00 AM to 11:20 AM
Speakers:
|
 |
Summary: |
Survey Propagation is an algorithm for random instances of constraint
satisfaction problems. It is derived from the cavity method for the
1-step replica symmetry breaking ansatz in spin glass systems.
We give an alternative view of the algorithm in the case of the SAT
problem. We describe how a SAT formula can be associated with a
novel family of Markov random fields (MRFs), parameterized by a real
number $\rho \in [0,1]$. We then show that applying belief
propagation--a well-known ``message-passing'' technique--to this family of MRFs
recovers various algorithms, ranging from pure survey propagation at one extreme
($\rho = 1$) to standard belief propagation on the uniform distribution
over SAT assignments at the other extreme ($\rho = 0$).
We further discuss experimental and theoretical result arising from the
study of the Markov Random Fields.
Joint work with Elchanan Mossel and Martin Wainwright. |
Keywords: |
Phase Transition;Computation;Reconstruction |
|
|
Lecture #10871
Need help? Visit our help pages at http://www.msri.org/communications/vmath/hints
|
|
|
|
| See more of our Streaming Videos on our main VMath - Streaming Video page. |