AI generates interpretation videos

AI extracts and analyses the key points of the paper to generate videos automatically Go Generating

AI Traceability

AI parses the academic lineage of this thesis Generate MRT

AI Insight

AI extracts a summary of this paper

Weibo:
We provide a simple characterization of it for tree structured graphs, and show how it can be used for approximations in non-tree graphs

An LP View of the M-best MAP problem.

NIPS, pp.567-575, (2009)

Cited by: 82|Views249
EI
Full Text
Bibtex
Weibo

Abstract

We consider the problem of finding the M assignments with maximum probability in a probabilistic graphical model. We show how this problem can be formulated as a linear program (LP) on a particular polytope. We prove that, for tree graphs (and junction trees in general), this polytope has a particularly simple form and differs from the ma...More

Code:

Data:

0
Introduction
• A common task in probabilistic modeling is finding the assignment with maximum probability given a model.
• Of particular interest is the case of MAP in graphical models, i.e., models where the probability factors into a product over small subsets of variables
• For general models, this is an NP-hard problem , and approximation algorithms are required.
• One possible approach is to use loopy max-product to obtain approximate max-marginals and use those to approximate the M best solutions 
• This is largely a heuristic and does not provide any guarantees in terms of optimality certificates or bounds on the optimal values.
Highlights
• A common task in probabilistic modeling is finding the assignment with maximum probability given a model
• The goal of the current work is to leverage the power of linear program (LP) relaxations to the M best case
• We begin by focusing on the problem of finding the second best solution. We show how it can be formulated as an LP over a polytope we call the “assignment-excluding marginal polytope”
• We approximated each MAP with its LP relaxation, so that both STRIPES and Nilsson come with certificates of optimality when their LP solutions are integral
• We provide a simple characterization of it for tree structured graphs, and show how it can be used for approximations in non-tree graphs
Methods
• The authors compared the performance of STRIPES to the BMMF algorithm  and the Lawler/Nilsson algorithm [6, 8].
• Nilsson’s algorithm is equivalent to PES where the 2nd best assignment is obtained from maximizations within O(n) partitions, so that its runtime is O(n) times the cost of finding a single MAP.
• The authors approximated each MAP with its LP relaxation, so that both STRIPES and Nilsson come with certificates of optimality when their LP solutions are integral.
• BMMF relies on loopy BP to approximate the M best solutions.7.
• For run-time comparisons, the authors normalized the times by the longest-running algorithm for each example
Conclusion
• The authors present a novel combinatorial object M (G, z) and show its utility in obtaining the M best MAP assignments.
• As with the marginal polytope, many interesting questions arise about the properties of M (G, z).
• In which non-tree cases can the authors provide a compact characterization.
• Another compelling question is in which problems the spanning tree inequalities are provably optimal.
• An interesting generalization of the method is to predict diverse solutions satisfying some local measure of “distance” from each other, e.g., as in 
Reference
Author  Amir Globerson 