How to use symbolism in machine learning?

How a matrix gets converted to an image I am not sure Symbolism and worked shape and size You want me when you’ve been specified So here I’ve got to India Cross-training It is the original emission is the dimensionally reducing it Now if we are not happy with this or let us say if you are making ah implementation where when the war security guard legacy runs the device through some of the costs they will get the number plate automatically or by looking at the low they will get the make off the car or by looking at the car will get the color of anything here You don’t need HD image.


You don’t have to compute or spend a lot of time and money and hardware under this review the image and then compute optics This could be a very simple application Okay so this is your argument But there is one more dimension reduction technique I’m not sure if you guys a court about it Have you heard about PS any No I don’t supervise learning No Yes I know No Right Okay Let me know if you guys are interested I would show something on this also in the power, Of course, Anything But I feel it’s little importance on times So let me show you first of all why we need to Yes any and very in all in deep learning you will need it so something of the blacker Not not in order The part, Of course, This is what we do in the industry I don’t want us to miss it See I have morning number.


It is a part of NLP So I show something short Something extracted some of my matches Now if you look at computer vision we saw today water transfer learning and didn’t that is vtg Imagine it’s election A leaner same thing You can do it in your NLP als where the honoree house some kind off including available ready-made So the first including that we can see here is glove Another one could be um you say Elmo another one could be worked awake So where do it is developed by Google So these are standard embodies that we have will directly important and start using is what we mean by plans for doing so again It’s an example So what if I say that in our text So what is this case story about As we have got a huge text available in us Now I want to know that uh let us say uh I want to predict something So I will say the prince to raise friends will be tomorrow’s question mark.


What we can do out of this I will say today it is friends is tomorrow’s Nash What is it So obviously you guys well know Prince is nothing But tomorrow he’ll be taking What if I remove the prince and said Today’s princess or today’s ah King Queen will be tomorrow’s what we don’t know So this is used when you want to do some kind of prediction Now they the best example I’ll give you is linguine So when you are messaging when Lyndon we have observed now So for example if Parrish sends Minnis message on Lyndon something so it’ll give me some kind off tokens.


If you’re seen it it will first say hi Harish for the morning Harish Sure Depending on the data are what is a return under If you also protect me sure will meet tomorrow or shoulder I will do that or something How do we get these predictions Use something called Caroline Okay Usar intend to do this now what is Iran in its a different network Recurrent Your little will do in NLP anyway but are in order to uses or we can sometimes or to use Club Glover’s, Ah the way you did Regional in a CV in NLP we can say ready-made a network So this is called Global Victor’s face and it was developed by Stanford University Now what in this What I’ll do is I’ll show you What if I want to predict what is the nearest according to king So from our corpus remember What is corpus I assure you over chat board corpus is a set-off huge data available to play in your model Okay now what if I say give me the top five words 1 to 6 minutes Five words which are the very new thing.


If you observe Queen Monarch Prince Kingdom and rain from our corpus can find these other five predictions Which one you want me to since I will say share the top three predictions always something like a recommendation system against for text All right now why do I am I’m showing you this because here if I want to represent this what if this this This is all back And I’m sure you basil them Have our district borders their data What is it What is he talking about If I want to represent this what if I present my tokens from my corpus able to visualize this Not very good but you can see that Yes There are some words we took tightly packed That means they’re very near to each other There are some words which are almost.


If you can order a university in college okay School and education hospital and health You can observe that these words are tagged together Now if you ask me how that’s again A different topic How I tack them together but glove will help me to do it Okay so how I represent Did this how to relate Convert Ah higher dimension imaged were to the image like this We use PSN You’re there So let me go through TSA What if he is sending a very simplistic way It is a supervised learning dimensional reduction technique So far we have been unsupervised learning dimension direction technique That is specie right In this case what will do us We will say there is a true data and we have got four points they want do 34 I will say this point Nobody at this point That would be point Embassy point,, somebody.


I will send a distance between again like me not answer Now I will ask you ways What are the similarity measures that you have seen so far Based on how I can find similarities Distance time similarity and distance walks immediately Okay In this case let me say that I’m saying distance would be one of the similarity matrices So I will say for our convenience I will say each one of them is that a unit distant from each other and diagonal elements will be year Distance elsewhere or cough two from each agreed Fight of Austria.


Now how would I use the diamond ship Let me ask you I want to reduce the dimension to one the house Right So to do that Okay let me not ask you Let me show you something Then I’ll ask you to do that when I will lose And a couple of point of reference Let us sail Start with an I’ll put a wager now next to a what do we have been our study We have to to be here at that unit distance and we have been it a unit distance And now if I say their distance between a B and B So now we are kind off close If I keep distance in my mind and if I replicate them on a one day space I can transform them like this.


Please remember how I might transport I’m transporting I’m saying that I will make sure that I don’t change their distances And right now the similarity matrix between all of them is distance So I have made in distances between a B and B agreed Everybody agrees to that Now what about where should I place my seat What if I place it here Because if you remember the distance between DNC it’s one And the distance between E and C is also squared off too So if I place it here if this is C what is gonna happen B and C is going to be far away right As BNC going is kind of one in Sofia So what if I bless See here Now it’s satisfied But now the will be an issue So is anyway satisfied with both of them but DnB air creating some issue Percy.

Leave a Comment