Big Data Analytics to Overcome “Swimming in Sensors and Drowning in Data” phenomenon: BIG DATA Analytics: Deriving Meaning/Insight from Data Explosion

Critical Infrastructure protection via Anomaly Detection

We have developed core technology that is based upon processing high dimensional data via diffusion processes, diffusion bases, diffusion geometries and other methodologies for finding meaningful geometric descriptions that represent normal behavior of the data and then identify deviations from normality in CI and networking data. We devise via diffusion geometries methodologies a profile (cluster/s) that uniquely describe/s the normal behavior of the data.

The main core of the methodology is based upon training the system to extract heterogeneous features, identify their profile that represents normal data behavior and then find patterns, which did not participate in the training, that deviate geometrically from the profile. These deviations are the malware anomalies we are after. This methodology offers behavioral analysis of heterogeneous complex networks via the geometry of the training data to produce a unified threat manager to maintain and preserve a networks' health.

It is universal generic core technologies for anomaly detections algorithms that fit different CI utilities, communication networks include cellular, SCADA and cloud computing that are based on well-founded deep unification between different mathematical theories from different disciplines that emerged in the last couple of years with classical mathematics such as applied and computational harmonic analysis, differential geometry, stochastic processing and classical analysis.

Project manager

Oren Calfon
VP Business Development, ICT

Project researchers

Amir Averbuch
T.A.U Tel Aviv University, Exact Sciences
School of Computer Science