Novel Algorithms and Software for Mapping Complex Networks

Complex networks are ubiquitous in both nature and engineered systems, including computer integrated circuits, the Internet, neurons in the brain, computational elements in artificial neural networks, data mining, and social networks.

Real world spatially and temporally defined complex networks that are able to represent, store, and process information are at the heart of a wide range of engineering disciplines. Advances in understanding the complexity of networks, and the ability to engineer ever more sophisticated ones, will contribute to both civilian and military applications, including faster more efficient computers and advances in biology and healthcare.

To create functional networks it is critical to understand quantitatively how nodes in complex networks are dynamically linked. Until now mapping the dynamic structure of geometrically defined complex networks with single node resolution has not been possible.

Technology Description:
This technology represents a method for mapping and analyzing the functional structure and topology of networked systems. The novel algorithms and software developed by UCSD can dynamically map the structure of complex networks with single node resolution. The algorithms can derive the topology of all the main classes of known networks in a high throughput way, independent of the networks’ physical details, and are optimized for real world spatiotemporal networks.

Benefits:

  • Unique computational tools that allow the analysis of a complex networked systems, providing an evolving map of the dynamic topology with single node resolution
  • Detailed quantitative information about the signaling behaviors and flow of information in the mapped complex network
  • Potential development of new computer software and systems with a variety of applications

Features:

  • Output of the algorithms is a dynamic map of the network topology as a function of both space and time
  • Algorithms map the network structure from a series of images recorded over time by qualitative observing the activation and de-activation of nodes in a network without any prior knowledge of the connectivity of the network.
  • Algorithm tested and validated using simulated and real experiment data
  • Potential development of single software or multiple software products

Market Potential/Applications :
There is a very significant commercial potential to contribute in a unique way to advanced engineering and medical applications. The ability to dynamically map the topology of physically different complex networks with single node resolution will have both civilian and military applications, including:

  • data mining and search engine technology
  • development of computer software, firmware, hardware, digital electronic circuitry, processors, data servers, etc.
  • understanding of functional information signaling in biological neural networks
  • bioinformatics (e.g. human genome project)
  • advance for understanding neurological and other cell signaling disorders
  • development of efficacy profiles for new drugs
  • identifying novel biomarkers for the diagnosis of diseases
  • testing physiological responses to external perturbations (e.g. disease mediating stressors)
  • development of novel artificial networks

Development Status:
This technology is offered exclusively or nonexclusively in the US and/or worldwide territories. A commercial sponsor for potential future research is sought.

Researcher:
Dr. Gabriel A. Silva is an Assistant Professor at UCSD, Department of Bioengineering, Jacobs School of Engineering and Department of Ophthalmology, School of Medicine.

Relevant Papers and Links:
http://www.silva.ucsd.edu/

Hashemi M, Buibas M, Silva GA.  Automated Detection of Intercellular Signaling in Astrocyte Networks Using the Converging Squares Algorithm. J. Neurosci Methods.  2008 Jan 29.

Case No: SD2006-258

Keywords: complex, network, artificial, algorithm, software, computer, program, engineering, robotics, node, structure, topology.

Inquiries To:  invent@ucsd.edu