BIGraph (Big Graph) is a research group within the EECS department at the Colorado School of Mines focused on modeling and analytics over large networks and graphs. Our collective expertise and interests within the focus area are broad and range from theory and algorithms to practical implementations and systems.
Our theoretical and algorithmic expertise includes combinatorial and spectral graph theory, game theory, graph algorithm design and randomized and approximation algorithms; our analytics expertise consists of machine learning and data mining including probabilistic graphical models; our systems expertise includes distributed systems, high performance computing including compiler-level optimization, GPUs and cloud computing. Application areas we are actively exploring include resource allocation in networks, sensor and robotic networks, mobile social networks, cheminformatics, bioinformatics, the smart grid, and computational materials.
Big data is a broad term for data sets so large or complex that traditional data processing applications are inadequate. In business applications, insights obtained from big data can lead to better decisions and improved outcomes; in science, the ability to analyze massive experimental data sets enables the discovery of new scientific facts and laws; and in engineering, smart large-scale data analysis helps to build more intelligent systems.
Within the larger big data space, BIGraph focuses on graph (or networked) data, where relationships between objects yield greater insights than studying them in isolation. Well-known large networked applications include computing PageRank on billions of web pages (think Google searches), social network analytics on hundreds of millions of people, path-finding on road networks containing hundreds of millions of road segments (think Google maps or GPS route guidance), electrical grid networks, oil and gas pipeline networks and many others.
Network analytics has its technical basis in graph theory, a branch of mathematics. Within Computer Science, fast algorithms that exploit the theory have been developed for a variety of graph applications over several decades. By mapping relationships among high volumes of highly connected data, graph analytics unlocks more insightful questions and produces more accurate outcomes. More recently, the computing community has been actively addressing the challenge of large network analytics by developing hardware and software platforms that are specifically designed to work with graph data, that is often stored in specialized graph databases. These include graph-processsing frameworks for distributed platforms (including the cloud), shared memory platforms, GPUs, etc.
The BIGraph research group is engaged with all of these aspects of large graph analytics along with applications that are built on an underlying network structures.
