Scribd
Upload
19 views

Lec 9

The document summarizes the Hopfield network, a type of neural network developed by John Hopfield. The Hopfield network is a fully interconnected network with symmetric weights and no self-connections. It has feedback between the output and input layers, so the present output is dependent on both the present input and previous outputs, giving it memory-like properties. The response of each neuron is determined by calculating the weighted sum of its inputs and applying an activation function. Examples are provided to demonstrate how the network updates its activations iteratively until convergence.

Uploaded by

mohammed.riad.bi.2020
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
19 views

Lec 9

The document summarizes the Hopfield network, a type of neural network developed by John Hopfield. The Hopfield network is a fully interconnected network with symmetric weights and no self-connections. It has feedback between the output and input layers, so the present output is dependent on both the present input and previous outputs, giving it memory-like properties. The response of each neuron is determined by calculating the weighted sum of its inputs and applying an activation function. Examples are provided to demonstrate how the network updates its activations iteratively until convergence.

Uploaded by

mohammed.riad.bi.2020
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/ 16

Part 3 :

Neural
Networks (NN)
Lecture 9

Presented by
Zainab Al-Qudsi
Objectives

❑The Hopfield Network

❑Example

2
3.2 The Hopfield Network

The Nobel prize winner (in physics ) John Hopfield has


developed the discrete Hopfield net in (1982-1984). The net
is a fully interconnected neural net, in the sense that each
unit is connected to every other unit. The discrete Hopfield
net has symmetric weights with no self-connections, i.e,

In this NN, inputs of 0 or 1 are usually used, but the weights


are initially calculated after converting the inputs to -1 or +1
respectively.

3
3.2 The Hopfield Network

4
3.2 The Hopfield Network

The outputs of the Hopfield are connected to the inputs as


shown in Figure, Thus feedback has been introduced into
the network. The present output pattern is no longer solely
dependent on the present inputs, but is also dependent
on the previous outputs. Therefore the network can be said
to have some sort of memory, also the Hopfield network has
only one layer of neurons.
The response of an individual neuron in the network is given
by :-

5
3.2 The Hopfield Network

The outputs of the Hopfield are connected to the inputs as


shown in Figure, Thus feedback has been introduced into
the network. The present output pattern is no longer solely
dependent on the present inputs, but is also dependent
on the previous outputs. Therefore the network can be said
to have some sort of memory, also the Hopfield network has
only one layer of neurons.
The response of an individual neuron in the network is given
by :-

6
3.2 The Hopfield Network

7
3.2 The Hopfield Network

8
3.2 The Hopfield Network

9
3.2 The Hopfield Network

10
3.2 The Hopfield Network

11
3.2 The Hopfield Network

12
3.2 The Hopfield Network

13
3.2 The Hopfield Network

2-The i/p vector x = (1 1 1 0). For this vector, y= (1 1 1 0)


Choose unit y1 to update its activation

14
3.2 The Hopfield Network

15
3.2 The Hopfield Network

3- Test for convergence, false


So that ,The input vector x = (1000), for this vector,
Y= (1 0 0 0)

16

You might also like