
Partial Optimality by Pruning for MAPInference with General Graphical Models
We consider the energy minimization problem for undirected graphical mod...
read it

Maximum A Posteriori Inference in SumProduct Networks
Sumproduct networks (SPNs) are a class of probabilistic graphical model...
read it

Exact MAP Inference by Avoiding Fractional Vertices
Given a graphical model, one essential problem is MAP inference, that is...
read it

MPLP++: Fast, Parallel Dual BlockCoordinate Ascent for Dense Graphical Models
Dense, discrete Graphical Models with pairwise potentials are a powerful...
read it

Exact MAPInference by Confining Combinatorial Search with LP Relaxation
We consider the MAPinference problem for graphical models, which is a v...
read it

CoarsetoFine Lifted MAP Inference in Computer Vision
There is a vast body of theoretical research on lifted inference in prob...
read it

Taxonomy of Dual BlockCoordinate Ascent Methods for Discrete Energy Minimization
We consider the maximumaposteriori inference problem in discrete graph...
read it
Maximum Persistency via Iterative Relaxed Inference with Graphical Models
We consider the NPhard problem of MAPinference for undirected discrete graphical models. We propose a polynomial time and practically efficient algorithm for finding a part of its optimal solution. Specifically, our algorithm marks some labels of the considered graphical model either as (i) optimal, meaning that they belong to all optimal solutions of the inference problem; (ii) nonoptimal if they provably do not belong to any solution. With access to an exact solver of a linear programming relaxation to the MAPinference problem, our algorithm marks the maximal possible (in a specified sense) number of labels. We also present a version of the algorithm, which has access to a suboptimal dual solver only and still can ensure the (non)optimality for the marked labels, although the overall number of the marked labels may decrease. We propose an efficient implementation, which runs in time comparable to a single run of a suboptimal dual solver. Our method is wellscalable and shows stateoftheart results on computational benchmarks from machine learning and computer vision.
READ FULL TEXT
Comments
There are no comments yet.