Transcript Application Of Evolutionary Neural Network Architecture

Application Of Evolutionary
Neural Network Architecture
Bharadwaj Krishnamurthy
Abstract
I am investigating the effectiveness of a new
evolutionary neural network model developed in
the Smart Engineering Systems Laboratory at
UMR by Amber Fischer. This model incorporates
modularity of functionality and the evolution of
exploratory mechanisms for the evolution of a
network of spiking neurons.
Current Work
• Right now I am building an architecture of
spiking neurons for recognition of sound signals.
The papers that I am referring to are:
1)http://neuron.princeton.edu/~moment/Organis
m/Docs/paper.pdf
2)http://neuron.princeton.edu/~moment/Organis
m/Docs/paperII.pdf
Current Work
Cont…
• Why am I doing this?
Once I build a simple neural network
architecture with spiking neurons on my own, I
will be in a better position to work with amber’s
evolutionary neural network architecture with
spiking neurons.
Current Work
Cont…
• I am also working on a paper which investigates
the effectiveness of Amber’s network in solving
group technology problems.
Abstract: Group technology has a number of
methods that seek to identify similarities
between the design of a product and the
manufacturing processes that are involved in its
production.
Current Work
Cont…
Abstract (Cont): The advantage of employing
group technology is that, that parts that
undergo similar manufacturing operations can
be grouped together, reducing machine setup
and down time. In this paper we are
investigating if the group technology problem
could be solved using biologically inspired
evolving neural networks
Future work
• Right now the evolutionary architecture
•
developed by amber is not integrated and is in
the form of different chapters. I am planning to
integrated all the chapters into one single
architecture.
Once this is done, I am planning to investigate
the effectiveness of this architecture in data
mining and other real time problems.
My Contact
• You can reach me at:
[email protected]