Skip Navigation
Search
Back to Search Innovations
Network Community Detection Based on Ricci Flow
Using Ricci flow and curvature to break down community groups within complex online networks

Source: Gerd Altmann, https://pixabay.com/photos/play-stone-network-networked-1237457/, Pixabay License.

Background

A majority of complex networks in the real world currently have groups of well-connected nodes that each have important functional roles, otherwise known as community structures. Using these community structures, a multitude of practical applications such as social network applications, biology and biochemistry applications, and Internet and Peer‑to‑Peer applications can be implemented.There is a shortage in the methods of community detection and identification that are efficient and accurate. Finding these communities are important as they give more information on the complex network and allow for a better analysis of it. Therefore, there is a need for a technology that can effectively identify communities within complex networks.

Technology

This technology revolves around community detections within complex networks by assigning the computation of the complex networks as geometric objects and assigning communities in a network as a geometric decomposition.Some embodiments of this process involve applying curvature and discrete Ricci flow to decompose smooth manifolds to break down communities in networks.By implementing this method and utilizing Ricci flow, it is more accurate and efficient at identifying communities that are behind the many layers of complex networks.

Advantages

There aren’t a lot of procedures to accurately and efficiently identify community structures within complex networks, so this technology provides that need. It allows for a more in‑depth analysis of complex networks via their individual communities.

Application

This technology will be used to identify certain community groups within any singular complex network.

Inventors

Jie Gao, Professor, Computer Science
Yu-Yao Lin, Dr., Computer Science
Chien-Chun Ni, Dr., Computer Science

Licensing Potential

Development partnerCommercial partnerLicensing

Licensing Status

Available for Licensing

Licensing Contact

James Martino, Licensing Specialist, OTLIR, james.martino@stonybrook.edu,

Patent Status

Provisional patent

62/911,644

Tech Id

050-9067