Department of Navy awards InfoBeyond the development of EStreaming that is a real-time graph data analytics that builds over COMAND collaborative machine learning platform for scalable graph embedding. It promotes collaborative analytical machine learning for processing big graph data in a real-time fashion. For example, it allows continuously monitor of a very large scale of network of data points and interacting edges. Unsupervised and semi-supervised machine learning algorithms can be conducted in EStreaming. It splits a huge graph into smaller subgraphs such that distributed graph embedding can be conducted in parallel among a set of processors. Meanwhile, the graph embedding can be effectively merged and visualized. It has the graph embedding analytics that are not provided by conventional graph embedding approaches:
· Continuously exploration of the graph internal structure to precisely reveal rich graph analytics (e.g., prediction of group split, statistic parameters such as closeness),
· Dynamic analysis to reveal the network evolution and capture the perishable events (e.g., identifying the similar or close partner, evaluating the impact of a new event, etc.),
· Discovery of new network knowledge on event detection, classification, tracking, etc.
EStreaming has the ability to discovery the spatial-temporal knowledge that did not discovered using other approaches.