Design and implementation of an artificial neuromolecular chip and its applications to pattern classification problems

  • Authors:
  • Yo-Hsien Lin;Jong-Chen Chen;Chao-Yi Huang

  • Affiliations:
  • Department of Information Management, National Yunlin University of Science and Technology, 123 University Road, Section 3, Douliou, Yunlin 64002, Taiwan, ROC;Department of Information Management, National Yunlin University of Science and Technology, 123 University Road, Section 3, Douliou, Yunlin 64002, Taiwan, ROC;Department of Information Management, National Yunlin University of Science and Technology, 123 University Road, Section 3, Douliou, Yunlin 64002, Taiwan, ROC

  • Venue:
  • Neurocomputing
  • Year:
  • 2009

Quantified Score

Hi-index 0.01

Visualization

Abstract

This paper describes the design and implementation of an artificial neuromolecular model, a biologically motivated architecture, with digital circuit. The artificial neuromolecular circuit (to be referred to as the ANM chip) is an evolvable hardware architecture that combines intra- and inter-neuronal levels of processing. Evolutionary learning algorithm is employed to train the chip for specific tasks. In this paper we applied it to pattern classification problems. Experimental results showed that the ANM chip was capable of learning in an autonomous manner. It also achieved a satisfactory result in pattern differentiation and noise tolerance. An interesting finding was that the chip enjoyed the synergy that percolated through different combinations of evolutionary learning operators.