A. Innovative claims for the proposed research (1)
1. Gathers a program distinctly different from other programs in taking account of the human component in a human/machine reasoning system and integrating and augmenting it for higher performance, in contrast to the conventional approach that attempts to create an autonomous reasoner with only supervisory participation by humans. This is an innovation in human information interaction, as recognized by NIMA. Our Knowledge Sharing Core architecture develops computational processes and data encoding that supports various types of human memory and anticipation. The Knowledge Sharing Core employs scientific concepts not widely understood or used as the basis for software. A synthetic cognition system (cognition understood to include perception) will be developed that reflects a natural stratification in human cognition and is interactive with it. Mathematical convolutions over localized bits of information, in the form of (type:value) pairs residing in the computer, are used to move categorical information from one level to a higher level. The school of ecological psychology traces how these levels are elaborated in interaction with natural environments. Project advisors Kugler and Shaw are pioneers in this field. Lakoff's philosophical argument for embodied realism (Lakoff, 1999) suggests that the approach we are using is not a small matter but is part of a scientific revolution, potentially overturning "first generation" cognitive science and software and the analytic philosophy from which it stems.
2. We have integrated patented processes from several sources, and in this integration process we are creating several new patents based on the input from scholars and from practical engineering experience. Because these patents are widely disclosed, this creates greater opportunity to understand the technology and find improvements, leading to more patents. A system that is fully disclosed avoids the blockage to scientific discovery and wide participation created by proprietary institutional structures that are hidden as part of business decisions. The Knowledge Sharing Core concept retains the opportunity for creative individuals to be rewarded, which is problematic in other open source programs not centered on patents. The result is a highly open, participatory process with a highly efficient reward mechanism, which results in much higher rates of innovation and performance.
3. Several advances in ontology processing, including data reduction using categorical abstraction into localized containers. Containers are created in those places where ambiguation or disambiguation processes are essential. The contents of these containers are sets of (type:value) pairs, ie CCM constructions, that are double encoded into hash tables to provide almost instantaneous set theoretic operations, including support for convolutions based on a type differentiating kernel operant. The tradition of multi-logics (von Forester) and quasi-axiomatic theory (Finn and Pospelov) is to be extended and used to supply inferential support for various processes, including automation of ambiguation and disambiguation during natural language parsing as part of Semio and NdCore ontology construction. Stochastic processes (from Recommind Inc) and latent semantic analysis (from SAIC) are to be used to reveal linguistic variations that may point to memetic expression in social discourse.
4. Uses multiple methods. The Knowledge Sharing Core concept changes the emphasis from saying whether one is better, to recognition that simultaneous use of processes, that are radically different, will perform better than any one method. As in nature, variety and selection are used as an aid to evolution of a domain specific extraction system.
5. Use and further development of an unencumbered infrastructure more conducive to ontology operations. Conventional infrastructure is optimized for both transaction processing and retention of proprietary structures. The Knowledge Sharing Core is optimized to support human knowledge sharing.