Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


ucin1122071899.pdf (4.79 MB)

ETD Abstract Container

Abstract Header

A COMPONENT LIBRARY OF ARTIFICIAL NEURON ACTIVATION FUNCTIONS

ATHREYA, JOOTHIRAM JAYAMANI
http://rave.ohiolink.edu/etdc/view?acc_num=ucin1122071899

Abstract Details

2005, MS, University of Cincinnati, Engineering : Electrical Engineering.
Hybrid Intelligent Systems (HIS) combine intelligent techniques in synergistic architectures to provide solutions to complex problems. They utilize at least two of the three techniques: fuzzy logic, genetic algorithms and neural networks. Example Systems have many applications: neurocomputers, industrial process control, data validation, risk management, texture analysis, 3-D object recognition, signal processing, and biomedical systems. Most of these applications are currently software based. Hardware implementations for certain of these applications are not only more accurate, but are also faster, more compact and can be made application specific. But since most of these applications require analog components, implementing them in hardware is costly and time-consuming. However, Hardware Description Languages provide a powerful medium with enhanced flexibility to model these components. Though the models based on the behavioral description of these components developed using the HDL technique can be easily validated, to validate the structural models we need a library of these components in hardware. In this research we have designed and characterized a library of some basic components like the Unit Step, Linear Threshold, Sigmoid and Gaussian functions in MAGIC to facilitate the validation of corresponding structural models of these functions developed in Verilog-AMS and VHDL-AMS. Although schematics for these components are widely available, for an actual working circuit we need complete characterization and layout, such as we have provided here, to support the HDL descriptions. The idea here is to facilitate rapid prototyping, a design technique that enables the use of basic building blocks available in a standard library for a wide range of applications.
Carla Purdy (Advisor)
138 p.

Recommended Citations

Citations

  • ATHREYA, J. J. (2005). A COMPONENT LIBRARY OF ARTIFICIAL NEURON ACTIVATION FUNCTIONS [Master's thesis, University of Cincinnati]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1122071899

    APA Style (7th edition)

  • ATHREYA, JOOTHIRAM. A COMPONENT LIBRARY OF ARTIFICIAL NEURON ACTIVATION FUNCTIONS. 2005. University of Cincinnati, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ucin1122071899.

    MLA Style (8th edition)

  • ATHREYA, JOOTHIRAM. "A COMPONENT LIBRARY OF ARTIFICIAL NEURON ACTIVATION FUNCTIONS." Master's thesis, University of Cincinnati, 2005. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1122071899

    Chicago Manual of Style (17th edition)

ucin1122071899
1,147
© 2005, all rights reserved.
This open access ETD is published by University of Cincinnati and OhioLINK.