CORA : A Flexible Hybrid Approach to Building Cognitive Systems

The Cognitive Reasoning and Representation Architecture (CORA) is a flexible hybrid approach and toolset that enables human-like reasoning across a variety of problem domains. It combines multiple existing reasoning techniques and implementations along with a framework to facilitate sharing of knowledge among reasoning modules, including (but not limited to) case-based reasoning (CBR), Bayesian networks, and fuzzy reasoning. CORA enables the exchange of various knowledge types across reasoning modules via point-to-point, point-to-multipoint, and shared-memory communications mechanisms. Three problem domains are demonstrated: cognitive wireless communications, electronic sensing for novel radar waveforms, and intelligence surveillance and reconnaissance (ISR) sensor retasking. CORA’s multi-platform implementation supports a variety of software languages to enable reasoning processes to take place using the best tools for each reasoning process, including the integration of several well-known open-source reasoning tools.