Modified Bifurcating Neuron with leaky-integrate-and-fire model

  • Authors:

  • Lon Risinger;Khosrow Kaikhah

  • Affiliations:

  • Department of Computer Science, Texas State University, San Marcos, Texas;Department of Computer Science, Texas State University, San Marcos, Texas

  • Venue:
  • IEA/AIE'2004 Proceedings of the 17th international conference on Innovations in applied artificial intelligence
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

The Modified Bifurcating Neuron (MBN) is a neuron model that is capable of amplitude-to-phase conversion and volume-holographic memory. Inputs are real valued and temporally spaced. This allows information to be coded in the temporal spacing of inputs and outputs as well as their values. At its core, the MBN incorporates a stateful leaky-integrate-and-fire neuron model. The MBN attempts to produce these properties by simulating mechanisms present in biological neural systems to a greater extent than is normally found in artificial neural networks. MBNs use an object model rather than the normal linear algebra approach. The MBN is conceptually based on the computational model presented in the "Bifurcating Neuron Network 2" by G. Lee and N. Farhat.