Transcript pre02

Dynamically Reconfigurable
Neurons
This presentation summaries the
progression achieved up to date.
• Artificial Neural Networks
• Implementing the ANNs on FPAA using
Anadigm Designer Software
• Literature Review
– Switching Capacitor
– PAMA Paper
This presentation summaries the
progression achieved up to date.
• Figuring out Pin-outs
• MUX Board Testing
– Dealing with NIDAQ
– Dealing with Amplicon DIO
Artificial Neural Network
Inputs
Neuron Body
x1
w1
Activation
Block
w1x1
x2
w2
w2x2
∑
∑wixi
y
Φ(x)
Output
xn
wn
wnxn
 n

y  f   xi wi 
 i 1

Artificial Neural Network
Input
SumDiff
x
g1
g2
Σ
Integrator
With Reset
reset
Output
c
∫
+
vref
y
Artificial Neural Network
Simulation of an SNN
FPAA – Anadigm Designer
FPAA – Anadigm Designer
FPAA – Anadigm Designer
FPAA – Anadigm Designer
FPAA – Anadigm Designer
Literature Review
• Switched Capacitor (SC) is a circuit made
of one capacitor and two switches which
connect the capacitor with a given
frequency alternately to the input an output
of the SC. This simulates the behaviors of
a resistor, so SCs are used in integrated
circuits instead of resistors. The resistance
is set by the frequency.
• PAMA Paper
Pin Outs
• Figuring out
– MUX Board pin-outs
– Base Board pin-outs
– Amplicon Connector pin-outs
MUX Board Testing
• Dealing with NIDAQ
– Installing NIDAQ Software
– Playing with NI own example code
– Searching on the net for more suitable code
which is easier to understand and easy to
apply
– Programming NIDAQ, i.e Reading and Writing
Voltages Successfully
MUX Board Testing
• Dealing with Amplicon PCI 272 Device
– PC Setup, Software Installation
– Working through the VB code to control the
card
– Use of Amplicon Device in the Project
• Address bits and WR line bits
Plans for Next 2 Months
• Implementing the MUX Boards to the
FPAA
– To be completed by mid-February
• Implementation of MUX Board with ANNs
– To be completed by the end of February
– Writing the GA code to establish the Dynamic
connections among neurons
• Writing the GA code to establish the
Dynamic connections among neurons