Emerging Software Algorithms for Autonomous Sense Making Operations

ABSTRACT: We propose an elegant network-building algorithm for sense-making applications capable of learning the spatial-temporal structure of applied input data. Through a combination of synaptic and structural plasticity our networks will dissect vast quantities of streaming information down to its natural building blocks and then spontaneously assemble a network structure that models the spatial-temporal structure of the information flowing into it, allowing the network to make predictions over time. By monitoring the internal changes in network structure our algorithm will detect anomalous activity. BENEFIT: Our sense-making algorithm will find use in applications needing analysis of spatial-temporal data for suspicious(anomalous) events. Examples of use include recognizing suspicious changes in a computer users activity on secured networks, anomaly detection in multi-spectral video, and classification of spatial-temporal patterns.