WEBVTT 00:00.360 --> 00:02.280 -: Hello, and welcome to the intuition tutorials 00:02.280 --> 00:06.060 for the Artificial Neural Networks part of the course. 00:06.060 --> 00:08.970 Super excited to get these things started 00:08.970 --> 00:10.020 and today we're going to find out 00:10.020 --> 00:11.850 how we're going to tackle this section. 00:11.850 --> 00:14.730 So in this section, we will learn the following things. 00:14.730 --> 00:16.440 First of all, we'll talk about the neuron. 00:16.440 --> 00:19.800 So, there'll be a little bit of a neuroscience 00:19.800 --> 00:23.160 and we'll find out a bit about how the human brain works 00:23.160 --> 00:25.740 and why we are trying to replicate that. 00:25.740 --> 00:28.890 And we'll also see what the main building block 00:28.890 --> 00:32.130 of a neural network of the neuron looks like. 00:32.130 --> 00:33.750 Then in the next tutorial we'll talk about 00:33.750 --> 00:35.370 the activation function, 00:35.370 --> 00:37.800 and we'll look at a couple of examples 00:37.800 --> 00:39.750 of activation functions that you could use 00:39.750 --> 00:41.040 in your neural networks, 00:41.040 --> 00:45.090 and we'll find out which one of them 00:45.090 --> 00:49.200 is the most commonly used one in neural networks 00:49.200 --> 00:53.820 and in which layers you would rather use which functions. 00:53.820 --> 00:56.430 Then we'll talk about how neural networks work. 00:56.430 --> 00:59.670 So, in contrast to what you would expect 00:59.670 --> 01:01.050 and what it was probably conveyed 01:01.050 --> 01:04.230 in other courses and tutorials, 01:04.230 --> 01:07.380 we're not going to go into the learning, 01:07.380 --> 01:09.750 we're actually going to go into the working 01:09.750 --> 01:11.580 of the neural networks first 01:11.580 --> 01:15.660 because that way, by seeing a neural network in action, 01:15.660 --> 01:19.860 that will allow us to understand what we're aiming towards, 01:19.860 --> 01:20.760 what our goal is. 01:20.760 --> 01:23.610 So here we'll look at an example of a neural network. 01:23.610 --> 01:26.790 We're going to look at a very simplified, 01:26.790 --> 01:28.620 a very simplified hypothetical example 01:28.620 --> 01:30.570 of a neural network working, 01:30.570 --> 01:32.610 to predict housing prices. 01:32.610 --> 01:34.980 So basically, real estate prices. 01:34.980 --> 01:37.890 And by looking at that example, we'll understand better 01:37.890 --> 01:40.050 exactly what we're aiming towards 01:40.050 --> 01:42.030 and what we want to achieve in the end. 01:42.030 --> 01:45.150 And then we will move on to understanding 01:45.150 --> 01:46.620 how neural networks learn 01:46.620 --> 01:51.240 because that way we'll be more prepared for what's coming. 01:51.240 --> 01:53.040 Then we'll talk about gradient descent. 01:53.040 --> 01:56.610 This is also part of neural networks learning 01:56.610 --> 01:59.160 and we'll understand how that algorithm 01:59.160 --> 02:02.490 is better than just the brute force method 02:02.490 --> 02:05.250 that you might be intending, 02:05.250 --> 02:08.160 or willing to take as a first resort, 02:08.160 --> 02:10.110 or first method that comes to mind. 02:10.110 --> 02:12.060 So, we'll find out how, 02:12.060 --> 02:14.490 what the advantage of gradient descent are. 02:14.490 --> 02:17.880 And then we'll talk about stochastic gradient descent. 02:17.880 --> 02:20.670 It's a continuation of the gradient descent tutorial, 02:20.670 --> 02:23.880 but it's an even better and even stronger method 02:23.880 --> 02:26.190 and we'll find out exactly how it works. 02:26.190 --> 02:27.600 And finally, we'll wrap things up 02:27.600 --> 02:31.740 by mentioning the important things about back propagation 02:31.740 --> 02:35.580 and summarizing everything in a step by step 02:35.580 --> 02:36.780 set of instructions 02:36.780 --> 02:40.410 for running your artificial neural networks. 02:40.410 --> 02:42.270 I hope this all sounds very exciting to you 02:42.270 --> 02:44.580 because I am very excited myself 02:44.580 --> 02:46.950 and I can't wait to get started. 02:46.950 --> 02:47.970 I look forward to seeing you 02:47.970 --> 02:50.370 on the first tutorial, and until then, 02:50.370 --> 02:51.603 enjoy deep learning.