A Unifying Framework for Hybrid Information Processing

摘要

In this paper we report preliminary results of our ongoing research on the development of a sound framework for processing simultaneously disparate data. Specifically, the space of primaries (SP) is introduced and delineated herein. A concrete primary is defined out of a probability space, where solvable equations between equivalence classes are treated in the corresponding metric lattice. A scheme for reasoning is proposed tentatively, that is the ABC-scheme, which can both (i) learn adaptively associations of antecedent to consequent events and (ii) generalize. We discuss how the space of primaries (SP) might come handy in dealing with such practical problems as reducing the number of rules in a fuzzy system, or developing dynamic mathematical models for various "intelligent" processes.