1 00:00:00,300 --> 00:00:08,010 Here we can show this single neuron with some inputs, with a set of inputs, which we can just represented 2 00:00:08,010 --> 00:00:14,410 with a vector of X, and then each neuron has its own way, a specific way. 3 00:00:14,430 --> 00:00:16,720 We can just show them with a vector of W. 4 00:00:16,950 --> 00:00:23,730 And there is also a constant variable V, which we call it bias of our system. 5 00:00:24,030 --> 00:00:26,820 And then finally, this one will give us some output. 6 00:00:27,270 --> 00:00:31,350 We have also some function, which we call it a function of net care. 7 00:00:31,400 --> 00:00:40,270 We have a function of net which equals W one, X one because of our inputs. 8 00:00:40,300 --> 00:00:41,880 It's just the picture of inputs. 9 00:00:41,880 --> 00:00:45,930 So they can be several inputs based on baseline of our data. 10 00:00:47,190 --> 00:00:56,160 W one x1 plus W two x2 plus let's say W off and. 11 00:00:58,160 --> 00:00:58,790 Xin. 12 00:01:02,160 --> 00:01:10,230 And of course, we have a constant number B now, we can surely win metric, you can say that these 13 00:01:10,230 --> 00:01:26,220 are W one, W two to W Alphen, and then here we have another matrix, which is X one, X two up to 14 00:01:26,220 --> 00:01:26,760 X of. 15 00:01:29,500 --> 00:01:30,190 Plus. 16 00:01:31,500 --> 00:01:42,330 Will meaning here we have horizontal metrics multiply by vertical metrics and in metrics, we know that 17 00:01:42,330 --> 00:01:51,950 this one would be X one times W one W two times X two and finally W and times X of it. 18 00:01:53,580 --> 00:02:04,410 And here we already know that our X is a vector of X, we can just shoot, we need X one x one X to 19 00:02:05,040 --> 00:02:16,020 up to here X and the same with overweights of our W which shows how effective is some input and what 20 00:02:16,020 --> 00:02:18,630 is the effect of each input on our output. 21 00:02:19,140 --> 00:02:30,780 W one W two to w off and we can already write it down as a W transpose because we have here to transpose 22 00:02:30,780 --> 00:02:33,180 this one is horizontal. 23 00:02:33,180 --> 00:02:42,310 Whatever matrix is a vertical so that we transpose X plus B and overall two, which is easy here. 24 00:02:42,420 --> 00:03:00,000 Now we can write it as Z equals F of net and here this net signal would be F, W, transpose X plus 25 00:03:00,300 --> 00:03:11,070 B, we just showed our output of this Neron by using a function of W transform, which is our weight 26 00:03:11,070 --> 00:03:14,090 vector and X which is our input. 27 00:03:14,100 --> 00:03:24,270 And finally, plus a constant number with a simple mathematical function we can represent over Narron. 28 00:03:25,200 --> 00:03:31,290 Let's take a look at this and get and see how can we just show that when a new formula that we just 29 00:03:31,290 --> 00:03:42,120 found out here we have and this is MASN y axis and X axis, these are over imputes and these circles 30 00:03:42,120 --> 00:03:46,890 are representing zero for output or let's say totally false in these. 31 00:03:46,890 --> 00:03:50,040 Drongo is showing one for outward or totally wrong. 32 00:03:50,280 --> 00:03:54,200 And we already learned that we can separate them with a simple line. 33 00:03:54,210 --> 00:03:58,300 This is an equation for this line and each one is representing one. 34 00:03:58,650 --> 00:04:03,810 Let's take a look at this equation one more time. 35 00:04:04,530 --> 00:04:06,690 We can have two output. 36 00:04:07,260 --> 00:04:14,010 Z is equal to here, X and Y, and it can have two output. 37 00:04:14,010 --> 00:04:16,030 This is either one or zero. 38 00:04:16,500 --> 00:04:26,540 This is one whenever X plus Y minus one point five, greater or equal than zero. 39 00:04:26,550 --> 00:04:30,630 Here is part these are above one. 40 00:04:30,630 --> 00:04:40,560 And for this one, the Z can have output of zero whenever X plus Y minus one point five is less than 41 00:04:40,560 --> 00:04:41,100 zero. 42 00:04:42,000 --> 00:04:47,860 But we're looking at this function and by looking at this equation, what can you remember reading? 43 00:04:48,240 --> 00:04:49,440 This is pretty clear. 44 00:04:49,440 --> 00:04:55,620 This is over a step function graph of T, which has two output, one or zero. 45 00:04:55,860 --> 00:04:58,690 Whenever a cheese grater equal, then zero. 46 00:04:58,710 --> 00:04:59,460 This is one. 47 00:04:59,460 --> 00:05:02,930 And whenever T is less than zero, that's zero. 48 00:05:03,510 --> 00:05:12,720 So meaning we can write it already Z equal to F of T and here our T is the net function. 49 00:05:14,310 --> 00:05:22,670 So we can also write it as F of X plus Y minus one point five. 50 00:05:23,640 --> 00:05:31,050 So meaning eliminate magical view, we can simply show and Narron with this function.