a) Define Kohonen networks with examples.
Kohonen networks, also known as Self-Organizing Maps (SOMs), are a type of artificial neural network developed by Teuvo Kohonen in the 1980s. These networks are particularly useful for unsupervised learning tasks, where the goal is to detect patterns and structures in data without explicit supervision.
In Kohonen networks, neurons are organized in a two-dimensional grid, where each neuron is associated with a weight vector of the same dimensionality as the input data. During training, the network adjusts its weights to map ___ ________ ______ _____ ____ ____ ________ ______ _______ ___ ______.
______ ______ __________ ___ _____ ____ __________ ______ ________ _________ ____.
_________ _____ _____ ___ _________ _______.
____ ___ __________ ____ ______ ___ _______ _________ ______ ________ ____.
________ ________ ____ ________ ________.
_____ ____ __________ ________ ___ _____.
______ __________ _____ ______ __________.
______ ________ _____ ________ _____.
_________ _____ _____ ______ ______ _____.
______ ____ _________ ___ ______ __________ ___ ________ _____ ___ _________.
_______ _______ __________ ___ _________ ____.
____ ___ _______ ______ __________ ______ _______ __________ __________ _____ _______ ________.
_______ _________ _____ ___ __________ ___.
____ _____ _____ ___ _____ _________ _____ _________.
___ ______ ________ _______ __________ ________ __________ ________ ___ ______ _____ ___.
______ ________ _________ ________ ___ __________ _____ ______ _______ __________ __________ __________.
_______ __________ ___ ______ _______ ________ ______ ___.
______ ___ ______ _________ ______ ________ _____ ________.
_____ __________ _________ _______ ____ ___ ________ ____ ________.
______ ________ _____ ______ ___ ____ ____ ______ ______ ________ ____.
_____ _____ _____ ___ ____.
_________ ______ _____ _____ _______ ________ _____ _______ _________.
_______ ____ _____ _______ ______ _______ _________ __________ _________.
_____ ______ ________ _______ ______ ______ ______ _________ ______ ____ __________ _____.
______ ________ _____ ______ ________ ________ _______.
_____ ______ _______ ________ ____ ____ _________ ____ ___.
________ _________ ________ ___ ______ __________.
_______ ________ ____ _________ ______ ________ __________ ________ _______.
___ _________ _____ ________ ________ _________ _______ ______ ________ ______ ________.
Get Full Answer on WhatsApp
Which of the following statements are true or false? Give a short proof or a counter example in support of your answers.
a) There is chance of occurrence of the premature convergence in Roulett-wheel selection shceme used in GA.
b) Gradient based optimization methods are used when the objective function is not smooth and one needs efficient local optimization.
c) The cut of a fuzzy set A in U is defined as A
d) A single perception with preprocessing is neither an auto associative network nor a multiple layer neural network.
e) If W then perception is non-linear separable.
b) Consider the following single layer perception as shown in the following figure.
and the activation function of each unit is defined as
Calculate the output y of the unit for each of the following input patterns:
| Patterns | ||||
| 1 | 0 | 1 | 1 | |
| 0 | 1 | 0 | 1 | |
| 0 | 1 | 1 | 1 |
b) Let an activation function be defined as
Show that What is the value of
at the origin? Also, find the value of
as v approaches
and
b) Apply the “very” hedge on the fuzzy sets defined in Q. 1(b) to get the new modified fuzzy sets. Show the modified fuzzy sets through numeration.
b) Describe the Function Approximation in MLP. Also, explain Generalization of MLP.
b) Write the types of Neural Memory Models. Also, give one example of each.
Consider a dataset of six points given in the following table, each of which has two features and
Assuming the values of the parameters c and m as 2 and the initial cluster centers
and
apply FCm algorithm to find the new cluster center after one iteration.
| 3 | 11 | |
| 3 | 10 |
| 8 | 12 | |
| 10 | 6 | |
| 13 | 6 | |
| 13 | 5 |