DARPA Ensemble-Based Modeling Large Graphs & Applications to Social Networks.

摘要

Economies, social and political systems all exist embedded within complex and dynamic networks. Understanding the network landscape is a critical necessity for winning conflicts and securing global safety, peace and stability. It helps prevent surprises and provides both a tactical and a strategic upper hand in our interactions both with our allies and our adversaries. Understanding complex networks requires the development of reliable mathematical and computational tools that efficiently probe the data and extract the relevant information. This is a very difficult undertaking and it requires marshaling methods from traditionally disparate areas, including graph theory (discrete mathematics), statistical physics, statistics, theoretical computer science, algorithm development, and data mining in a sustained fashion. This project has been developing mathematical and computational methods pushing the state of the art in data driven modeling of complex networks. The algorithms and the methods developed here have been validated against real-world network datasets.