Dynamic Binding Communication Mechanism
Penn collection
Degree type
Discipline
Subject
Funder
Grant number
License
Copyright date
Distributor
Related resources
Author
Contributor
Abstract
Shastri & Ajjanagadde have proposed a biologically plausible connectionist rule-based reasoning system (hereafter referred to as a knowledge base, or KB), that represents a dynamic binding as the simultaneous, or in-phase, activity of the appropriate nodes [9]. This paper makes the first attempt at designing a biologically plausible connectionist interface mechanism between 2 distinct phase-based KB, as the next step toward providing a computational account of common-sense reasoning. The Dynamic Binding Communication Mechanism (DBCM) extracts a dynamic binding from a source KB and incorporates the binding into a destination KB so that it is consistent with the knowledge already represented in the latter. DBCM consists of several distinct, special-purpose modules. The Binding Memory (BM) is made up of several identical banks of nodes. Each time a temporally-encoded dynamic binding is extracted from the source KB, it is transferred into one of the banks, where the binding is converted to a spatially-encoded representation. The Phase Database (PD) monitors the target KB and produces a phased output that is out of phase with all other nodes in the target KB. The Phase Allocator (PA) synthesizes information from the Phase Database and from the target KB to determine the phase in which to introduce the new dynamic binding into the target KB. In turn, the PA extracts a single binding from one of the banks in the BM and introduces it into the target KB. The interface also utilizes 2 searchlight mechanisms: the first governs which bank in the BM receives bindings; the second mediates between the active banks (those which are currently representing bindings), and the Phase Allocator.