Question

Consider the multi-layer perceptron whose architecture is defined by three matrixes of synaptic weights with the following dimensions: W₁ E R6 x 8, W₂ E R4x7, W3 ER3 X5

a) How many features are applied to the MLP?
b) What is the number of neurons in each layer n₁, n₂, n3?
c) What is the number of classes?

Answer 1

The multi-layer perceptron has 8 features, 6 neurons in the first hidden layer, 4 neurons in the second hidden layer, and 5 classes.

Step-by-step explanation:

a) The number of features applied to the MLP is equal to the number of columns in the input matrix. The first weight matrix W₁ has dimensions 6x8, so it represents 8 features.

b) The number of neurons in each layer can be found by looking at the dimensions of the weight matrices. We have n₁=6 neurons in the first hidden layer, n₂=4 neurons in the second hidden layer, and n₃=3 neurons in the output layer.

c) The number of classes can be determined by looking at the dimensions of the last weight matrix. W₃ has dimensions 3x5, so it represents 5 classes.