ENSIGN is cutting-edge hypergraph analysis technology for Big Data spanning security, finance, geospatial applications, and biology.
ENSIGN is a multi-aspect analysis tool for Big Data.
ENSIGN models hypergraphs as sparse tensors (higher dimensional arrays), and leverages sparse tensor decomposition algorithms to subject the data to a spectral analysis. It is analogous to shining the data through a prism to see what constitutes it, via analyzing its spectrum.
ENSIGN is an advanced form of graph analytics; the use of higher dimensional arrays and multi-linear algebra allows semantics beyond connectivity to be incorporated as first class entities in the spectral analysis. For example, time, type, temperature, place, activity, channel, protocol, and many other attributes become part of the graph analysis. The hypergraph form allows joining different kinds of data (e.g., genomic + proteomic) for systems analysis. Working in higher dimensions, ENSIGN can find activities that may not be explicitly connected, but are linked through correlations in these other dimensions.
It is well known that linear algebraic approaches to graph analysis can provide amazing insights. For example, the famous PageRank algorithm, the keystone of Google, is about finding the first eigenvector of the web connectivity graph to find and rank the important of pages. Sparse tensor analysis is about taking this to higher dimensions, using multi-linear algebra, to bring in semantics.
While the use of sparse tensor decomposition has existed for decades, these algorithms were not practical for large scale data analysis due to the complexity of the computations. Reservoir’s patented algorithms and data structures provide breakthrough capabilities so that ENSIGN can process very large hypergraphs into interesting components.
ENSIGN makes it easier to identify anomalies and enables “on the fly” analysis of streaming data, thus shortening response times in critical cybersecurity situations.
Additionally, ENSIGN’s ability to analyze large complex patterns and monitor streaming activities makes it a valuable tool in finance applications.
Similarly, in biology, ENSIGN can be used to gain new insights by analyzing DNA sequences in the presence of other factors such as individual, species, or time.
In geospatial applications, ENSIGN can provide insights into patterns of life and aid the detection of anomalous behavior.
ENSIGN is provided as a software package that can be installed on Linux systems ranging in capacity from laptops to supercomputers. It includes graphical APIs that to support analyst preparation of data and visualization of results, as well as interfaces (e.g., Python) that allow ENSIGN to be integrated into other analytics flows.
ENSIGN provides a unique range of different types of sparse tensor decompositions, streaming tensor decomposition capability, and scale, to enable new kinds of Big Data analysis, beyond the capability of other tools.