A Temporal Sequence Processor Based on the Biological Reaction-diffusion Process
Sylvian R. Ray
Hillol Kargupta
Department of Computer Science,
University of Illinois at Urbana-Champaign,
Urbana, IL 61801, USA
Abstract
Temporal sequences are a fundamental form of information and communication in both natural and engineered systems.The biological control process that directs the generation of iterative structures from undifferentiated tissue is a type of temporal sequential process. A quantitative explanation of this temporal process is reaction-diffusion, initially proposed in [1] and later widely studied and elaborated.
We have adapted the reaction-diffusion mechanism to create a temporal sequence processor (TSP) composed of cells in a diffusion-supporting medium that performs storage, associative retrieval, and prediction for temporal sequences. The TSP has several interesting attributes: (1) Its achievable depth (or degree) is constrained only by storage capacity, (2) it tolerates substantial time warp, and (3) it supports user-specified flexible groupings of stored sequences into coarser classes at retrieval time. The TSP is also capable of preserving the time extent of stored symbols, as in a musical melody, and permits retrieval of both the symbols and their temporal extent. Experimental verification of the properties of the TSP was performed with Reber grammar sentences and musical melodies.
