Regionalized Optimization
This work provides a foundational mathematical framework for optimization problems with multiple perspectives on datasets, potentially impacting various domains in machine learning and AI.
The authors introduced a theoretical framework for constructing a global loss from local losses under functor constraints, called regionalized loss, and derived associated message passing algorithms for optimization. They demonstrated that Generalized Belief Propagation algorithms fit within this framework and proposed new algorithms for noisy channel networks.
We propose a theoretical framework for non redundant reconstruction of a global loss from a collection of local ones under constraints given by a functor; we call this loss the regionalized loss in honor to Yedidia, Freeman, Weiss' celebrated article `Constructing free-energy approximations and generalized belief propagation algorithms' where a first example of regionalized loss, for entropy and the marginal functor, is built. We show how one can associate to these regionalized losses message passing algorithms for finding their critical points. It is a natural mathematical framework for optimization problems where there are multiple points of views on a dataset and replaces message passing algorithms as canonical ways of finding the optima of these problems. We explain how Generalized Belief propagation algorithms fall into the framework we propose and propose novel message passing algorithms for noisy channel networks.