Scribd
Upload
14 views2 pages

Unit 5

Data mining

Uploaded by

sharveshram120
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
14 views2 pages

Unit 5

Data mining

Uploaded by

sharveshram120
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

UNIT-V

1. Define a dynamical system in the context of neurodynamics.


2. Explain the concept of stability of equilibrium states. How do you
determine if an equilibrium state is stable?
3. What are attractors, and how do they function within a dynamical system?
4. Differentiate between fixed point attractors and limit cycle attractors with
examples.
5. Explain how the manipulation of attractors applies to recurrent networks.
6. Describe the architecture of a Hopfield network. How is it different from
feedforward networks?
7. Explain how the energy function is used in Hopfield models. How does it
contribute to the network’s ability to settle into attractor states?
8. Discuss how atypical attractors can be introduced or manipulated as a
method for improving the memorization capacity of a Hopfield network.
9. How is energy minimization related to solving optimization problems with
Hopfield networks? Formulate a simple optimization problem (e.g., a
small instance of the Traveling Salesman Problem or a binary decision
problem) that can be potentially solved using a Hopfield network model.
10. Select a neurodynamical model of your choice and describe how digital
simulation tools could be used to study its dynamics, stability, and
attractor properties.
11. List three practical applications for neurodynamical systems and Hopfield
networks. Explain how the concepts and models are applied in these
cases.
12. Consider a simple dynamical system defined by the differential equation
( -x + x3 )
a) Find the equilibrium points of this system.
b) Analyze the stability of each equilibrium point using the linearization
method.

13. Consider a Hopfield network with three neurons. The connection weights
are given by the following symmetric weight matrix:
W = 0 1 -1
1 0 1
-1 1 0
If the initial state of the network is {S} = [1, -1, 1], update the network
asynchronously one neuron at a time. Which attractor state does the network
settle into?

14. For a Hopfield network with weights specified as wij = si sj ; Verify using a
Hebbian learning rule that a given pattern p = [1, -1, 1, 1] remains an
attractor.

15. For a Hopfield network with two neurons, the weight matrix is given by:
W=0 3
3 0
i. Find the eigenvalues of this weight matrix.
ii. Using the eigenvalues, determine if the network is likely to have stable
equilibrium points.

16. A simple neural system is described by the equation: dx/dy = -3x + x2


i. Find the equilibrium points.
ii. Use a Lyapunov function candidate V(x) = x2 to analyze the stability
of each equilibrium point.

16. A neural network’s activity is governed by the equations:


x(t) = 2 - 2e{-0.5t}
y(t) = 3 - 3e{-0.5t}
i. Calculate the values of x(t) and y(t) at t = 2 and t = 4.
ii. Determine if the system is converging to a point attractor and, if so, identify
its coordinates.
17. A two-neuron recurrent network has dynamics given by:
dx/dt = -x + ay
dy/dt = -y + bx
where a and b are parameters. Suppose a = 1 and b = 2.
i. Find the equilibrium points for this network.
ii. Determine the effect on the equilibrium points if the parameter a is
increased to 3.

You might also like