The EVS envisioned would dramatically improve the capability of human analysts and Internet and wireless appliances to make inferences in complex, data-intensive domains. The EVS would vastly increase the ability to identify individuals at point-of-entry border crossings, airports, and other locations warranting the validation of their identity. The system would search for key facts hidden in immense quantities of irrelevant information and assemble large numbers of disparate facts to reach valid conclusions in real time (see Figure 11.1).
Figure 11.1: Incremental profiles are distributed.
The system would be capable of learning in the following ways:
Learn from the data: make effective use of a wide variety of public records from government and commercial databases
Learn actively: request new data and analyses that optimally improve learning and inference
Learn cumulatively: incrementally improve existing knowledge and make use of that knowledge in subsequent learning and inference
From humans and machine algorithms: make use of data mining techniques under the direction of knowledgeable analysts and investigators