1 00:00:02,070 --> 00:00:10,890 ‫Now we are going to begin our journey into the field of neural networks work on neural networks began 2 00:00:10,890 --> 00:00:18,870 ‫with the motivation that human brain computes in a very different way than a conventional computer and 3 00:00:18,930 --> 00:00:26,250 ‫human brain is able to perform certain types of computations such as image recognition and pattern recognition 4 00:00:26,640 --> 00:00:30,360 ‫much better than a conventional computer. 5 00:00:30,360 --> 00:00:37,860 ‫On the other hand computers are very good at complex arithmetic calculations which human brain is not 6 00:00:37,860 --> 00:00:41,420 ‫very good at. 7 00:00:41,590 --> 00:00:49,560 ‫Let us take an example of recognizing handed and digit recognizing so many different types of handwritten 8 00:00:49,560 --> 00:00:58,600 ‫digits is not easy but we humans are astoundingly good at making sense of these digits. 9 00:00:59,160 --> 00:01:06,480 ‫On the other hand if we have to write a computer program to decode did it like these can you imagine 10 00:01:06,480 --> 00:01:08,420 ‫how difficult it is. 11 00:01:08,730 --> 00:01:09,940 ‫What will be the. 12 00:01:10,110 --> 00:01:20,580 ‫To say that these two are the same digit nine and this one is a seven and not a nine. 13 00:01:21,090 --> 00:01:31,020 ‫When we try to make such precise rules we get stuck with a number of exceptions caveat and special cases. 14 00:01:31,020 --> 00:01:35,280 ‫This is where neural networks come in. 15 00:01:35,280 --> 00:01:39,450 ‫Neural networks do not have the inbuilt rules. 16 00:01:39,450 --> 00:01:48,360 ‫Rather neural networks infer these rules for recognizing handwriting from a given large set of handwritten 17 00:01:48,360 --> 00:01:53,250 ‫digits which is also known as the training examples. 18 00:01:53,250 --> 00:02:02,010 ‫So basically we will take a large set of data feed into the neural network neural network will make 19 00:02:02,190 --> 00:02:12,180 ‫some rules and we will use this brain neural network to predict for the other cases this ability to 20 00:02:12,180 --> 00:02:19,230 ‫create own rules makes neural network versatile powerful and scalable. 21 00:02:19,590 --> 00:02:24,030 ‫In this course we are going to take the example of fashion M NASD 22 00:02:27,770 --> 00:02:39,590 ‫fashion M NIST is a data set in which we classify images into N different fashion items in this dataset. 23 00:02:39,620 --> 00:02:48,610 ‫We will have images of several fashion items and we will classify them into 10 different categories. 24 00:02:48,650 --> 00:02:53,690 ‫For example the first image is offer ankle boot. 25 00:02:53,720 --> 00:02:56,210 ‫The second is a pullover. 26 00:02:56,240 --> 00:02:59,570 ‫It is a trouser and so on. 27 00:02:59,570 --> 00:03:07,310 ‫So we will have grayscale images of several fashion items and will classify them into several fashion 28 00:03:07,340 --> 00:03:07,880 ‫product 29 00:03:11,060 --> 00:03:16,850 ‫this problem statement is going to be significantly more challenging than the handwriting recognition 30 00:03:16,850 --> 00:03:18,410 ‫example. 31 00:03:18,410 --> 00:03:25,580 ‫And once you are done with this you will appreciate the power of Neural Networks. 32 00:03:25,580 --> 00:03:31,970 ‫Let me give you an overview of what the next few lectures are going to be. 33 00:03:32,120 --> 00:03:39,920 ‫We will start the conceptual part by understanding two very important artificial neural cells called 34 00:03:40,180 --> 00:03:43,760 ‫perception and a sigmoid neuron. 35 00:03:43,760 --> 00:03:50,270 ‫Then we will tagged these single cells into multiple cells and we will make a multi-level perception 36 00:03:50,390 --> 00:03:56,780 ‫on a multilevel perception is the most commonly used neural network model. 37 00:03:57,320 --> 00:04:02,230 ‫Then we will dig deeper and understand forward and backward propagation. 38 00:04:02,600 --> 00:04:08,960 ‫Then we move on to stochastic gradient descent which is the standard learning algorithm for Mueller 39 00:04:08,960 --> 00:04:17,060 ‫networks and throughout the course we will be implementing all these models in Python. 40 00:04:17,810 --> 00:04:23,180 ‫In this course I will provide you with the reason of why we are doing whatever we are doing. 41 00:04:23,540 --> 00:04:31,250 ‫So by the end of the course you will understand what neural network and deep learning is how to create 42 00:04:31,910 --> 00:04:38,180 ‫a neural network to solve complex problems and when and where to use neural networks.