00:00And basically, it updates the weights, the updation of the weights happen after the backpropagation
00:05algorithm and where we take the gradient of the loss with respect to the weights because
00:10weights are the deciding factor for this error.
00:14So, after that one we came to a numerical on feed forward neural network and we saw
00:22if the inputs are some numbers and we give a particular activation function like a sigmoid
00:30activation function, how my outputs are calculated.
00:33And then we have come to the deep neural network, sorry come to a shallow and deep neural networks
00:41and we have seen how they are different, what are the applications where we can use shallow
00:47and deep neural networks and how the approaches differ when it comes to an image classification
00:55problem.
00:57So that is a nutshell for today's lecture.
01:01I hope you have found it informative and in the next session we will explore in detail
01:11the activation and the loss functions and because they are an integral part of the multilayer
01:18perceptrons.
01:20As you have seen, we have used but we have not gone into the details in this session.
01:24We will do it in the next session which is on activation functions and loss functions.
01:29Till that time, happy learning.
Comments