On the Design of A Malleable Chip: A Biologically Motivated Evolvable Hardware

Computer hardware has high processing speed; however, it is lack of flexibility in dealing with most real world problems. Evolvable neuromolecular hardware, motivated from some biological evidence, which integrates inter- and intra-neuronal information processing, is proposed. Evolvability, self-organizing dynamics, and a close structure-function relationship are three biological features captured in this hardware. The Quartus II system, a tool developed by the Altera Inc. for simulating digital circuits, is used to test this hardware and perform a series of experiments. The hardware is applied to pattern recognition problem domain. Experimental result shows that the structures and functions of this hardware are closely related that its outputs change accordingly when we gradually modify the structure, implying this hardware embracing an adaptive fitness landscape that facilitates evolutionary learning