Perceptual Reasoning Managed Big Data Analytics and Information Fusion

摘要

The focus and the challenge herein was to model an adaptive big data processing and information fusion system emulating the human perceptual reasoning/cognitive functions to facilitate two-way interactions between the visual data display and the analysts. Methods were presented with proposed solutions, issues and challenges using a combination of Data Mining, Visual Analytics and Perceptual Reasoning/Cognitive processing algorithms. The proposed systems are to enable the analyst to both refine the displayed data and via Resource Manager (RM) tasking elicit additional information in order to optimize confidence in decision making by feedback control. The analyst's cognitive functions were modeled/emulated/aided by the Perceptual Reasoning Machine (PRM) paradigm, a meta-level information management system. Part of the PRM uses associative recall, which was shown implementable via Case-Based Reasoning. The issues and challenges in the BDA and Information fusion system chain coupled with the PRM paradigm imbedded within the associated information Process Module System (PMS) were addressed. The associated interaction between PRM (which also serves as model for high level fusion processing) and the fusion levels was illustrated. Applications include Cognitive Intent Modeling (JDL Levels 2/3) from Social Networks (SNs) enormous "tweets" data feeds was addressed. Another application of the PRM/PMS was illustrated to address the challenges of the new project, entitled "Complex Adaptive Systems and Networks". There are many issues and challenges remaining requiring research, implementation and testing of the proposed methods.