Half-Space and Box Constraints as NUV Priors: First Results
This work is incremental, extending prior NUV methods to include specific constraint types for applications in signal processing or optimization.
The paper tackled the problem of incorporating half-space and box constraints into linear Gaussian models by proposing NUV representations, enabling these constraints to be added without compromising computational tractability.
Normals with unknown variance (NUV) can represent many useful priors and blend well with Gaussian models and message passing algorithms. NUV representations of sparsifying priors have long been known, and NUV representations of binary (and M-level) priors have been proposed very recently. In this document, we propose NUV representations of half-space constraints and box constraints, which allows to add such constraints to any linear Gaussian model with any of the previously known NUV priors without affecting the computational tractability.