Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated: 1 00:00:00,420 --> 00:00:07,320 What Allison should know what are you saying I don't know what Internet is that massive computer. 2 00:00:07,380 --> 00:00:10,870 Was the one that's becoming really big now. 3 00:00:11,010 --> 00:00:11,400 What do you mean. 4 00:00:11,440 --> 00:00:16,350 That's just what you know what do you write to what like I don't know a lot of people use it and communicate. 5 00:00:16,410 --> 00:00:18,960 I guess they can communicate with NBC writers and producers. 6 00:00:18,960 --> 00:00:21,180 Allison can you explain what Internet is 7 00:00:29,140 --> 00:00:30,590 how amazing is that. 8 00:00:30,670 --> 00:00:34,470 Just over 20 years ago people didn't even know what the internet was. 9 00:00:34,510 --> 00:00:37,260 And today we can't even imagine our lives are for it. 10 00:00:37,300 --> 00:00:39,560 Welcome to the people earning ETAs that course. 11 00:00:39,580 --> 00:00:44,800 My name is Cora Menko and along with the contractor had Lunda Pontmercy were super excited to have you 12 00:00:44,800 --> 00:00:45,570 on board. 13 00:00:45,610 --> 00:00:51,670 And today we're going to give you a quick overview of what deploring it is and why it's picking up right 14 00:00:51,670 --> 00:00:52,300 now. 15 00:00:52,300 --> 00:00:53,590 So let's get started. 16 00:00:53,590 --> 00:00:57,640 Why did we have a look at that clip and what is this photo over here. 17 00:00:57,670 --> 00:01:00,310 Well that clip was from 1994. 18 00:01:00,310 --> 00:01:03,110 This is a photo of computer from 1980. 19 00:01:03,220 --> 00:01:09,490 And the reason why we kind of delving into history a little bit is because neural networks along with 20 00:01:09,490 --> 00:01:15,670 deep learning have been around for quite some time and they've only started picking up now and impacting 21 00:01:15,670 --> 00:01:16,640 the world right now. 22 00:01:16,750 --> 00:01:22,510 But if you look back at the 80s you'll see that even though they were invented in the 60s and 70s they 23 00:01:22,510 --> 00:01:30,760 really caught on to a trend or called the cold wind in the 80s so people are talking about them a lot. 24 00:01:30,760 --> 00:01:35,890 There was a lot of research in that area and everybody thought that deep learning or neural networks 25 00:01:35,890 --> 00:01:41,830 were this new thing that is going to impact the world is going to change everything is going to solve 26 00:01:41,860 --> 00:01:46,060 all the world problems and they did kind of slow they died off over the next decade. 27 00:01:46,080 --> 00:01:51,210 And so what happened why did why did the neural networks not survive and not change the world with it 28 00:01:51,200 --> 00:01:51,750 was it. 29 00:01:51,940 --> 00:01:57,220 The reason for that that they were just not good enough that they are not that good at predicting things 30 00:01:57,270 --> 00:02:02,320 are not that good at modeling and messy just not a good invention. 31 00:02:02,350 --> 00:02:03,400 Or is there another reason. 32 00:02:03,400 --> 00:02:08,410 Well actually there is another reason and the reason is in front of us it's the fact that technology 33 00:02:08,410 --> 00:02:15,640 back then was not up to the right standard to facilitate neural networks in order for neural networks 34 00:02:15,700 --> 00:02:17,130 and deep learning to work properly. 35 00:02:17,140 --> 00:02:21,880 You need two things you need data and you need a lot of data and you need processing power you need 36 00:02:21,880 --> 00:02:25,960 strong computers to process that data and facilitate and you know that works. 37 00:02:25,990 --> 00:02:32,900 So let's have a look at how as data or storage of data has evolved over the years and then we'll look 38 00:02:32,900 --> 00:02:34,880 at how technology has evolved. 39 00:02:34,880 --> 00:02:39,350 So here we go got three years 1956 1980 2017. 40 00:02:39,830 --> 00:02:43,220 How much did storage look back in 1956. 41 00:02:43,280 --> 00:02:47,600 Well there's a hard drive and that hard drive is only a five. 42 00:02:47,750 --> 00:02:54,980 Wait for a megabyte harddrive That's five megabytes right there on the forklift the size of a small 43 00:02:54,980 --> 00:03:01,040 room that's a hard drive being transported to another location on a plane. 44 00:03:01,370 --> 00:03:04,670 And that is what storage looked like in the. 45 00:03:04,700 --> 00:03:11,360 In 1956 you had to pay a company had to pay two and a half thousand dollars of those days dollars to 46 00:03:11,390 --> 00:03:16,400 rent that hard drive to rent it not buy it or rented for one month. 47 00:03:16,400 --> 00:03:18,760 In 1980 the situation improved a little bit. 48 00:03:18,800 --> 00:03:24,290 So here we got a 10 megabyte hard drive for three and a half thousand dollars is still very expensive 49 00:03:24,290 --> 00:03:27,260 and only 10 megabytes So that's like one photo these days. 50 00:03:27,260 --> 00:03:36,840 And today in 2017 we've got a 256 gigabyte SD card for $150 which can fit on your finger. 51 00:03:37,100 --> 00:03:43,790 And if you're watching this video a year later or like in 2019 or 2025 you probably laughing to us all 52 00:03:43,790 --> 00:03:47,240 because by then you have even stronger storage capacity. 53 00:03:47,240 --> 00:03:52,910 But nevertheless the point stands if we compare these across the board and we even taking price and 54 00:03:52,910 --> 00:03:58,370 size into consideration just the capacity of whatever was trending at the time. 55 00:03:58,370 --> 00:04:04,090 So from 1956 to 1980 capacity increased about double. 56 00:04:04,250 --> 00:04:12,380 And then it increased about twenty five thousand six hundred times and the know the length of the period 57 00:04:12,380 --> 00:04:20,210 is not that different from 1956 to 1980 24 years from 1980 to 2013 thirty seven years so not that much 58 00:04:20,210 --> 00:04:24,610 of an increase in time but a huge jump in technological progress. 59 00:04:24,890 --> 00:04:28,220 And that stands to show that this is not a linear trend. 60 00:04:28,220 --> 00:04:33,830 This is an exponential growth in technology and if we add into it take into account price and size you 61 00:04:33,910 --> 00:04:37,090 will be in the millions of increase. 62 00:04:37,280 --> 00:04:40,540 And here we actually have a chart on a logarithmic scale. 63 00:04:40,640 --> 00:04:46,290 So if we plot the harddrive cost per gigabyte you'll see that looks something like this. 64 00:04:46,430 --> 00:04:49,880 We're very quickly approaching zero. 65 00:04:49,910 --> 00:04:54,200 Right now you can get storage on Dropbox and Google Drive which doesn't cost you anything. 66 00:04:54,200 --> 00:05:01,280 Cloud storage and that's going to continue and in fact over the years this is going to go even further. 67 00:05:01,280 --> 00:05:05,860 Right now scientists are looking into using DNA for storage. 68 00:05:05,960 --> 00:05:13,720 And right now it's quite expensive it costs $7000 to synthesize two megabytes of data and then another 69 00:05:13,730 --> 00:05:15,210 thought 2004's to read it. 70 00:05:15,230 --> 00:05:19,670 But that kind of reminds you of this whole situation of the harddrive and the plane you know that this 71 00:05:19,670 --> 00:05:24,590 is going to be mitigated very very quickly with this exponential curve 10 to 10 years from now 20 years 72 00:05:24,590 --> 00:05:28,590 from now everybody's going to be using DNA storage if we go down this direction. 73 00:05:28,660 --> 00:05:33,770 And here are some stats on that so you can explore it further maybe pause the pause the video if you 74 00:05:33,770 --> 00:05:35,360 want to read a bit more about this. 75 00:05:35,360 --> 00:05:36,890 This is from nature dot com. 76 00:05:37,040 --> 00:05:44,990 And basically you can store all of the world's data in just one kilo one kilogram of DNA storage or 77 00:05:44,990 --> 00:05:49,350 you can store about 1 billion terabytes of data in one gram of DNA storage. 78 00:05:49,350 --> 00:05:56,150 So that's just something to to show how quickly we're progressing and that this is why deep learning 79 00:05:56,150 --> 00:06:02,840 is picking up now that we are finally at the stage where we have enough data to train super cool super 80 00:06:02,840 --> 00:06:04,280 sophisticated models. 81 00:06:04,280 --> 00:06:08,350 Back then in the 80s when I was first initially invented juice just wasn't the case. 82 00:06:08,690 --> 00:06:15,740 And the second thing we talked about is processing capacity So here we've got an exponential curve again 83 00:06:16,460 --> 00:06:17,850 on a log scale. 84 00:06:17,850 --> 00:06:23,840 It's an ideally portrayed here but on the right because it's a log scale and this is how computers have 85 00:06:23,840 --> 00:06:24,410 been evolving. 86 00:06:24,410 --> 00:06:30,530 So again feel free to post the slide this is called Moore's Law you've probably heard of it how quickly 87 00:06:30,530 --> 00:06:34,120 the processing capacity of computers has been evolving. 88 00:06:34,310 --> 00:06:39,920 Right now we're somewhere over here where an average computer can buy for a thousand bucks thinks at 89 00:06:39,920 --> 00:06:47,960 the speed of the brain of a rat and between two and five will be the speed of a human or 20:23 and then 90 00:06:48,470 --> 00:06:54,740 by 2050 or 2045 it will surpass all of the humans combined. 91 00:06:54,740 --> 00:07:01,610 So basically we're entering the era of computers that are extremely powerful that can process things 92 00:07:01,610 --> 00:07:05,670 WAY faster then then we then we can imagine. 93 00:07:05,720 --> 00:07:08,490 And that is what is facilitating the learning. 94 00:07:08,600 --> 00:07:14,510 So all of this brings us to the question What is deep learning what what is this whole neural network 95 00:07:14,510 --> 00:07:18,390 situation what what is going on what are we even talking about here. 96 00:07:18,410 --> 00:07:20,490 And you've probably seen a picture or something like this. 97 00:07:20,570 --> 00:07:21,590 So let's dive into it. 98 00:07:21,590 --> 00:07:29,140 What is deep learning this gentleman over here Jeffrey Hinton is known as the godfather of deep thing. 99 00:07:29,350 --> 00:07:37,030 And he did research on deep learning in the 80s and he's done lots and lots of work lots of research 100 00:07:37,030 --> 00:07:41,370 papers he's published in deep learning right now. 101 00:07:41,370 --> 00:07:42,920 He works at Google. 102 00:07:43,030 --> 00:07:47,770 So a lot of the things that we're going to be talking about actually come from Jeffrey Hinton and you 103 00:07:47,770 --> 00:07:48,260 can see a lot. 104 00:07:48,300 --> 00:07:49,840 He's got quite a few YouTube videos. 105 00:07:49,840 --> 00:07:53,930 He explains things really well so I highly recommend checking them out. 106 00:07:54,130 --> 00:07:59,140 And so the idea behind deep learning is to look at the human brain. 107 00:07:59,140 --> 00:08:01,710 And this guy is going to be quite a bit of neuroscience coming up. 108 00:08:01,750 --> 00:08:09,370 And in these tutorials and what we're trying to do here is to mimic how the human brain operates. 109 00:08:09,400 --> 00:08:13,300 And you know we don't know that much you don't know everything about the human brain but that little 110 00:08:13,300 --> 00:08:16,730 man that we all know we want to mimic it and recreate it. 111 00:08:16,730 --> 00:08:17,230 And why is that. 112 00:08:17,230 --> 00:08:21,970 Well because the human brain seems to be one of the most powerful tools on this planet for learning 113 00:08:22,300 --> 00:08:28,690 for learning adapting skills and then applying them and if computers could copy that then we could just 114 00:08:28,990 --> 00:08:32,980 leverage what natural selection has already decided for us. 115 00:08:32,980 --> 00:08:37,840 All of those kind of algorithms that it has decided are the best which are going to leverage that. 116 00:08:37,840 --> 00:08:39,750 Why reinvent the bicycle ride. 117 00:08:39,880 --> 00:08:41,700 So let's see how this works. 118 00:08:41,710 --> 00:08:50,260 Here we've got some neurons so these neurons which have been smeared onto glass and then have been looked 119 00:08:50,260 --> 00:08:52,210 at under a microscope with some coloring. 120 00:08:52,270 --> 00:08:57,310 And this is you can see what they look like so they have like a body they have these branches and they 121 00:08:57,310 --> 00:09:02,700 have like tails and so and so you can see them they have like a nucleus inside in the middle and that's 122 00:09:02,710 --> 00:09:06,940 that's basically what a neuron looks like in the human brain. 123 00:09:06,970 --> 00:09:12,310 There's approximately 100 billion neurons all together so these are individual neurons these are actually 124 00:09:12,310 --> 00:09:17,380 motor neurons because they're bigger they're easier to see but nevertheless there's a hundred billion 125 00:09:17,470 --> 00:09:19,740 neurons in the human brain. 126 00:09:20,000 --> 00:09:23,920 And it is connected to as many as about a thousand of its neighbors. 127 00:09:23,920 --> 00:09:26,590 So to give you a picture this is what it looks like. 128 00:09:26,590 --> 00:09:32,140 This is an actual data section of the human brain. 129 00:09:32,140 --> 00:09:38,950 And this is the cerebellum which is this part of your brain at the back. 130 00:09:38,950 --> 00:09:47,140 It is responsible for like more Torex and for you know keeping a balance and some language capabilities 131 00:09:47,140 --> 00:09:47,680 and something like that. 132 00:09:47,680 --> 00:09:57,460 So this is just to show how Vorst How many neurons there are like billions and billions and billions 133 00:09:57,460 --> 00:10:02,770 of neurons all connecting It's like we're talking about five or five hundred or a thousand or millions 134 00:10:02,770 --> 00:10:04,780 billions of neurons in there. 135 00:10:04,910 --> 00:10:08,350 And so that's that's what we're going to be trying to recreate. 136 00:10:08,350 --> 00:10:11,700 So how do we recreate this in a computer. 137 00:10:11,890 --> 00:10:20,200 Well we create an artificial structure called an artificial neural net where we have nodes or neurons 138 00:10:20,590 --> 00:10:26,500 and we're going to have some neurons for input value so these are values that you that you know about 139 00:10:26,500 --> 00:10:27,340 a certain situation. 140 00:10:27,340 --> 00:10:32,080 So for instance you're modeling something you want to predict something you always could have some input 141 00:10:32,080 --> 00:10:33,310 something to start. 142 00:10:33,310 --> 00:10:36,740 Your prediction is off then that's called the input layer. 143 00:10:36,820 --> 00:10:38,100 Then you have the output. 144 00:10:38,140 --> 00:10:43,780 So that's of value that you want to predict or it's surprise whether it's is somebody going to leave 145 00:10:43,930 --> 00:10:45,980 the bank or stay in the bank. 146 00:10:46,260 --> 00:10:50,530 Is this a fraudulent transaction it's a real transaction and so on. 147 00:10:50,890 --> 00:10:52,420 So that's going to be output lower. 148 00:10:52,540 --> 00:10:55,330 And in between we're going to have a hidden layer. 149 00:10:55,330 --> 00:11:01,600 So as you could see in your brain you have so many neurons so some information is coming in through 150 00:11:01,600 --> 00:11:04,990 your eyes ears nose so basically your senses. 151 00:11:05,140 --> 00:11:10,210 And then it's it's not just going right away to the output where you have the result is going through 152 00:11:10,240 --> 00:11:15,130 all of these billions and billions and billions of neurons before guess output and this is the whole 153 00:11:15,130 --> 00:11:17,050 concept behind it that we're going to model the brain. 154 00:11:17,050 --> 00:11:23,170 So we need these hidden layers that are there before the output so the input Lares neurons connected 155 00:11:23,170 --> 00:11:26,580 to a hidden layer neurons that neurons are connect to output. 156 00:11:26,880 --> 00:11:29,280 And so this is this is pretty cool. 157 00:11:29,290 --> 00:11:30,550 But what is this all about. 158 00:11:30,550 --> 00:11:34,040 Where is the deep learning here or why is it called deeper nothing deeper in here. 159 00:11:34,090 --> 00:11:41,090 While this is kind of like an option which one might call shallow learning where there isn't much indeed 160 00:11:41,110 --> 00:11:41,890 going on. 161 00:11:41,890 --> 00:11:43,400 But why is it called deploring. 162 00:11:43,420 --> 00:11:50,020 Well because then we take this to the next level we separate it even further and we have not just one 163 00:11:50,020 --> 00:11:58,180 hit and there we have lots and lots and lots of hidden layers and then we connect everything just like 164 00:11:58,180 --> 00:12:01,930 in the human brain connect everything interconnected everything. 165 00:12:01,930 --> 00:12:08,290 And that's how the input values are processed through all these hidden layers just like in the human 166 00:12:08,290 --> 00:12:08,660 brain. 167 00:12:08,770 --> 00:12:12,380 Then we have an output value and now we're talking deep learning. 168 00:12:12,460 --> 00:12:15,910 So that's what learning is all about on a very abstract level. 169 00:12:15,940 --> 00:12:21,180 And the further tutorials we're going to dissect and dive deep into deep learning and by the end of 170 00:12:21,180 --> 00:12:26,480 it you will know what the planning is all about and you will know how to apply it in your projects. 171 00:12:26,500 --> 00:12:31,910 Super excited about this wait to get started and I look forward to seeing the next tutorial. 172 00:12:31,950 --> 00:12:33,700 Until then enjoy deep learning. 19253