You are on page 1of 1

PART B

Answer all the questions.


Question 1
Convolutional net has an input of size m × n × 3 (i.e m is width and n is the height). The first layer is a
convolutional layer with f output feature maps and kernels of size k.

(a) Give a formula for the size of the output feature map with padding p and stride s? [5 marks]
(b) Give a formula for the number of learnable parameters in this layer (including biases) if we use
relu as the activation function? [5 marks]
(c) If you have a 32 × 32 × 3 image and filter it with 20 kernels of size 5 × 5 × 3, the way most convolu-
tional neural networks are implemented. If you use no padding and stride is 1, what will be the
size of the output feature map? and hence calculate the number of learnable parameters.
[10 marks]
(d) The figure Q1(d) depicts the part of convolutional neural network. Answer the following ques-
tions by referring to the figure.

Figure Q1(d): A part of convolutional neural network

i. Fill the values for the spaces labeled as A, B,C , D, E , F,G, H , I , J in the figure. [10 marks]
ii. Calculate the number of learnable parameters in the following Convolutional neural network
(including biases). Your answer should illustrate the all details of the calculations. [10 marks]

IT 5427 - Neuroscience and Neurocomputing Page 1 of 1

You might also like