AF: Small: Algorithms for Active Learning of Interaction Networks

Yale University, New Haven CT

Investigators

Abstract

The project will seek efficient algorithms for extracting the structure of interaction networks: systems consisting of finite populations of elements in which the state of each element may change as a result of interactions with a small set of other elements according to specific rules of interaction. Such networks are ubiquitous in the physical and social sciences, and include standard models such as Boolean circuits, Bayesian networks, social networks, chemical systems, gene regulation networks, and epidemiological models of the spread of disease. The research carried out will apply methods of active learning based on recent progress by the principal investigators on determining the structure of certain kinds of Boolean, analog and probabilistic circuits and social networks using experiments.