Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated: 1 1 00:00:00,300 --> 00:00:02,210 Hello and welcome to the Intuition tutorials 2 2 00:00:02,210 --> 00:00:05,960 for the artificial neural networks part of the course. 3 3 00:00:05,960 --> 00:00:08,870 Super excited to get these things started 4 4 00:00:08,870 --> 00:00:09,920 and today we're going to find out 5 5 00:00:09,920 --> 00:00:11,760 how we're going to tackle this section. 6 6 00:00:11,760 --> 00:00:14,650 So, in this section we will learn the following things. 7 7 00:00:14,650 --> 00:00:16,410 First of all we'll talk about the neurons 8 8 00:00:16,410 --> 00:00:19,680 so there'll be a little bit of neuroscience 9 9 00:00:19,680 --> 00:00:23,080 and we'll find out a bit about how the human brain works 10 10 00:00:23,080 --> 00:00:25,640 and why we're trying to replicate that 11 11 00:00:25,640 --> 00:00:28,800 and we'll also see what the main building block 12 12 00:00:28,800 --> 00:00:32,040 of a neural network, the neuron, looks like. 13 13 00:00:32,040 --> 00:00:33,660 Then in the next tutorial we'll talk about 14 14 00:00:33,660 --> 00:00:36,700 the activation function and we'll look at 15 15 00:00:36,700 --> 00:00:38,900 a couple of examples of activation functions 16 16 00:00:38,900 --> 00:00:41,090 that you could use in your neural networks 17 17 00:00:41,090 --> 00:00:43,680 and we'll find out which ones of 18 18 00:00:43,680 --> 00:00:45,890 which one of them is 19 19 00:00:45,890 --> 00:00:48,150 the most commonly used one 20 20 00:00:48,150 --> 00:00:49,840 in neural networks and 21 21 00:00:49,840 --> 00:00:53,720 in which layers you would rather use which functions. 22 22 00:00:53,720 --> 00:00:56,945 Then we'll talk about how neural networks work so 23 23 00:00:56,945 --> 00:00:59,530 in contrast to what you would expect 24 24 00:00:59,530 --> 00:01:01,740 and what was probably conveyed in other 25 25 00:01:02,790 --> 00:01:04,150 courses and tutorials, 26 26 00:01:04,150 --> 00:01:07,280 we're not going to go into the learning 27 27 00:01:07,280 --> 00:01:09,660 we're actually going to go into the working 28 28 00:01:09,660 --> 00:01:11,500 of the neural networks first. 29 29 00:01:11,500 --> 00:01:15,550 Because that way, by seeing a neural network in action 30 30 00:01:15,550 --> 00:01:18,920 that'll allow us to understand what we're 31 31 00:01:18,920 --> 00:01:20,660 aiming towards, what our goal is 32 32 00:01:20,660 --> 00:01:23,510 so here we'll look at an example of a neural network 33 33 00:01:23,510 --> 00:01:25,610 we're going to look at a 34 34 00:01:25,610 --> 00:01:28,550 a very simplified, very simplified hypothetical example 35 35 00:01:28,550 --> 00:01:30,500 of a neural network working 36 36 00:01:30,500 --> 00:01:32,510 to predict housing prices 37 37 00:01:32,510 --> 00:01:34,499 so basically real estate prices 38 38 00:01:34,499 --> 00:01:37,800 and by looking at that example we'll understand better 39 39 00:01:37,800 --> 00:01:39,950 exactly what we're aiming towards 40 40 00:01:39,950 --> 00:01:41,950 and what we want to achieve in the end. 41 41 00:01:41,950 --> 00:01:44,270 And then we will move on to 42 42 00:01:44,270 --> 00:01:46,510 understanding how neural networks learn 43 43 00:01:46,510 --> 00:01:51,130 because that way we'll be more prepared for what's coming. 44 44 00:01:51,130 --> 00:01:52,930 Then we'll talk about gradient descent. 45 45 00:01:52,930 --> 00:01:56,510 This is also part of neural networks' learning 46 46 00:01:56,510 --> 00:02:00,040 and we'll understand how that algorithm is 47 47 00:02:00,040 --> 00:02:02,410 better than just the brute force method 48 48 00:02:02,410 --> 00:02:04,010 that you might 49 49 00:02:04,010 --> 00:02:08,090 be intending or willing to take as a first resort 50 50 00:02:08,090 --> 00:02:10,810 or first method that comes to mind, so. 51 51 00:02:10,810 --> 00:02:12,900 We'll find out how great, what the advantage 52 52 00:02:12,900 --> 00:02:14,410 of gradient descent are. 53 53 00:02:14,410 --> 00:02:17,050 And then we'll talk about stochastic gradient descent. 54 54 00:02:17,050 --> 00:02:20,570 It's a continuation of the gradient descent tutorial 55 55 00:02:20,570 --> 00:02:23,780 but it's an even better and even stronger method 56 56 00:02:23,780 --> 00:02:26,080 and we'll find out exactly how it works. 57 57 00:02:26,080 --> 00:02:28,640 And finally we'll wrap things up by 58 58 00:02:28,640 --> 00:02:31,640 mentioning the important things about backpropagation 59 59 00:02:31,640 --> 00:02:35,500 and summarizing everything in a step-by-step 60 60 00:02:35,500 --> 00:02:37,520 set of instructions for running 61 61 00:02:37,520 --> 00:02:40,330 your artificial neural networks. 62 62 00:02:40,330 --> 00:02:42,190 I hope this all sounds very exciting to you 63 63 00:02:42,190 --> 00:02:44,460 because I am very excited myself, 64 64 00:02:44,460 --> 00:02:46,840 and I can't wait to get started. 65 65 00:02:46,840 --> 00:02:49,190 I look forward to seeing you on the first tutorial 66 66 00:02:49,190 --> 00:02:51,503 and until then, enjoy deep learning. 5715