Neuromolecularware - A bio-inspired evolvable hardware and its application to medical diagnosis

Abstract

Computer systems have powerful computational capability. However, it is brittle in that a slight program modification can inadvertently change the system functions. Biological systems demonstrate better adaptability than computer systems. An evolvable neuromolecular hardware motivated from some biological evidence is proposed. The hardware was further applied to medical diagnosis with a clinical database of premature babies who are given total parental nutrition (TPN). Experimental results show that the neuromolecular hardware was capable of learning to differentiate data in an autonomous manner. © 2008 Springer-Verlag Berlin Heidelberg.