Algorithms for Inference of Temporal Dynamics in Networks and Concentration of Measure Techniques for their Analysis

Research Project, 2012 – 2014

The goal of this proposal is to develop scalable algorithms that can be used to integrate dynamic data such as time series with the existing large collections of static data in order to understand how networks change dynamically in response to both external environmental stimuli, internal structural perturbations (such as gene knockouts) and how they evolve over time. While aiming to develop practical scalable algorithms that can be applied to real data, we also aim to ground our work in rigorous theoretical frameworks. Since a full rigorous analysis of the complex inference algorithms are currently beyond the reach of the state of the art, we aim to study simplified versions of our models and algorithms for rigorous theoretical analysis inspired by recent breakthroughs of this kind.