What is Artificial Intelligence | Artificial Intelligence Tutorial For Beginners | Edureka Rewind

Key Takeaways

hello everyone this is Sarah from Eddie Eureka and today me and reshma will be discussing what is artificial intelligence so we'll be looking at various applications of artificial intelligence we'll also see how to achieve artificial intelligence and finally we'll be discussing about the artificial intelligence course provided by director but first let us see is artificial intelligence a threat to human beings So speaking in that context saurabh tell me what comes to your mind when you think of a highly Advanced robot the picture that I have in my mind is a super high five Warrior robot fighting with Advanced Machinery embedded in its own body you know like a robot with a Canon arm and another with a machine gun arm like in Transformer movies okay but does it ever come to your mind sort of that what if such a powerful robot or that Autobot points again at you instead of the enemies because maybe it doesn't want to be controlled by you anymore when reshma that will actually scare me think about it if humans can create machines that are more intelligent than us then there is something to be worried about don't you think sorry yes of course and even when you see movies like these where robots aren't Savage the theory seems quite plausible doesn't it I completely agree with you reshma and let me tell you a very interesting story about this robot by Hanson robotics the name of this robot is Sophia so Sofia is pretty much similar to the one in the movie ex machina I believe you have seen that movie yes I have all right so she is able to behave like an adult with a learning of less than a year which takes humans 20 years to learn so I have a question for you reshma how much time it took for you to learn how to ride a bicycle I guess I was about 12 when I learned how to ride a bicycle and what if I tell you that Sophia can do it probably in few weeks so if our learning power is so robust I think she'll just need a couple of years to outsmart humans and she even said in an interview that she wants to destroy Humanity oh my and even Elon Musk I don't think so that guy needs any introduction I will just say that he's the most farsighted businessman I've ever seen and he has commented on Twitter that artificial intelligence is more dangerous than North Korea I'll just say that I'm not ready for a third world war and speaking the context of AI I think it becomes scary when a machine starts to do something that it wasn't supposed to do or it starts doing things unexpected so take this for an example the Google AI translation tool so as a part of its training it was taught to translate from English to Korean Korean to English then English to Japanese and Japanese to English but surprisingly after that it was directly able to translate from Japanese to Korean and Korean to Japanese without using English as a bridge and let me tell you sort of that it wasn't part of its training so it basically developed its own Intermediate Language to translate Japanese and Korean on its own you see what machines can do on their own and this is called zero shocked so that's pretty fascinating to know reshma but I don't think so we should scare our audience further and I think as long as we know how exactly it works things will be in our control so let us understand what is artificial intelligence sureshma I would like to know your thoughts about what is artificial intelligence because when I go on internet and type artificial intelligence I see multiple definitions I see people have given their own definitions about artificial intelligence so what are your thoughts about it I like to break down the term and then understand the meaning of it so you can see that the term artificial intelligence is composed of two words artificial and intelligence so artificial means anything man-made and it's just basic second standard definition an intelligent refers to intellect our thinking power which is an abstract thing and most of the times we think that things that we cannot see or the abstract things is not something that we can recreate but it is not the case we have created a man-made thinking power or intellect and that is what artificial intelligence is according to me but that completely makes sense Richmond I would also like to share my thoughts about it so artificial intelligence is basically creating intelligent softwares making a machine that is smart a machine that has a power to think analyze and make decisions based on that for example if I teach reshma how to multiply two by four I don't think so I need to teach her how to multiply four by two so that is nothing but intelligence so if you can put that intelligence inside a system that is called artificial intelligence that's a very nice definition sort of thank you reshma I'm glad you find it useful every artificial intelligent software is based on a particular goal and now you guys must be wondering how to achieve artificial intelligence so there are numerous ways of achieving artificial intelligence so let us discuss those approaches we will be discussing machine learning first let me give you a proper definition of that machine learning is a type of artificial intelligence that provides computers with the ability to learn without being explicitly programmed I know you didn't get anything guys so reshma why don't you simplify it for us so the best way of learning something is by taking a look at an example and there is an example in front of your screen so here we train our machine on various types of dog images and then we feed in a dog image on which the machine is already trained on and then it will recognize the dog in the image now let us discuss one more scenario so here are the image we feed in is of a dog which or machine is not trained on and because of that a machine won't be able to recognize that image so what do we do now sorrow sureshma the answer to this challenge lies inside your brain so this challenge was overcome by Deep learning which is inspired from how a human brain works you can even say that it uses neural networks so the challenge that we saw in the previous slide how our brain will solve that problem so when we see different types of dogs our brain tends to create an abstract image of the dog or you can say that we recognize certain features which are common to all the dogs like a dog's nose his ears tail far Etc so even if we have not seen that particular type of dog we can recognize it based on the general features well reshma that is nothing but deep learning in a nutshell so with deep learning we try to extract few features which are very general which are very abstract that you'll find in almost all the dogs that exist in the world and based on those features we can recognize the picture that we are feeding it it contains a dog all right so let me just give you a quick introduction to deep learning now so deep learning is a subfield of machine learning concerned with algorithms inspired by the structure and function of the brain called artificial neural networks now reshma please simplify this for us so here you see different layers the first layer is called the input layer then the two hidden layers and finally the output layer so these layers have multiple artificial neurons that is inspired by brain cells called neurons and in every layer some kind of function is performed the output of the first layer becomes the input to the next and we have separate tutorial on deep learning and neural networks so you can go through that for detailed explanation of deep learning we'll leave the links of those videos in the description box below so you can check it out from there now the question that I have in my mind reshmael but I think a lot of people must be thinking about it is artificial intelligence new well AI is actually perceived as a thing of a future or has a very high technology thing that has very recently become a buzzword but is it actually new so let us take a look at the history of artificial intelligence for that Cesaro wanted to take us through it serious man I'm pretty sure you'll be amazed to know that artificial intelligence was introduced by McCarthy in the year 1956 and around the 60s era when everyone were involved in cultural and political Evolution somewhere of their people were thinking about creating artificial now make us introduce this in his paper which was present in conference that was held mainly on the subject of artificial intelligence even Alan Turin proposes during tests in 1950s in a Turing test an interrogator asked questions and if we can't distinguish whether the answer is coming from a machine or human the machine passes the Turing test or you could say that machine is intelligent and guys this is happening in the 50s so reshma when your family was sending telegrams and there was no internet people were researching about artificial intelligence that's pretty fascinating to me it is pretty fascinating to me as well sort of and now when I'm thinking about this I think the concept of man-made beings dates back to the medieval times is it yes of course because there were mentions of mechanical men and artificial beings in Greek myths such as the golden robots of hephistas like Talos and pygmalions Galicia and there were also rumors and stories where people talked about placing a mind into matter and this is exactly the main concept behind artificial intelligence making a brain and placing it into a mechanical body so guys when people say that artificial intelligence is new or it's a thing of future you can hash them up with the facts like this and it's pretty fascinating you know I sometimes wonder reshma tension people were really able to do that well I don't know about ancient people sorrow but we sure can do it now and we are doing it now so let us take a look at the applications of AI to show you how much we have advanced in the field of AI and how much dependent are we on AI now so we are using AI in different fields there are several U.S Academy institutions that are employing artificial intelligence to tackle some of the world's greatest economic and social challenges they're using AI to determine the poverty struck areas in the world and determine ways to overcome it similarly even aviation industry reshma one such example will be data location and another example that I can think of is ticket price determination and reshma I have heard that robots are even teaching students Yes you heard it right sorry robots are even teaching people now they're known as intelligent tutoring system or its and there is an its called Sherlock that teaches Air Force technicians to diagnose electrical problems in aircrafts all right viewers I have a question for you what do you think reshma is a robot or she is actually a human being no I am not well I don't even think that robots speak like this anymore they're so human-like today so we'll leave our viewers wondering so comment if you think I'm a robot or not so let us know fine so let's move forward and talk about Healthcare so in healthcare with the help of artificial neural networks people are trying to Modern human body parts they're even analyzing the health history of a patient to predict what sort of diseases he's more prone to suffer from in the future now Richmond will tell us about how it can be used in heavy Industries so this is a field where we have to be dependent on artificial intelligence like in automobile manufacturing factories it becomes hazardous for humans to lift all those heavy machine parts in chemical factories we humans cannot work in a toxic environment and that is why we need a machine to do that for us and let me tell you that in 2014 China Japan the United States the Republic of Korea and Germany together amounted to 70 percent of the total sales volume of robots now I have a question for you Sarah do you think robots will take all our jobs in future if they're advancing so much well I'm not really sure about that reshma but I'll give you a very good example so 100 years back when people were talking about computers the same assumption was made for computers as well so they thought that computers will take all of our jobs but what's the scenario right now computer has generated so many jobs so I believe that's the same will be with artificial intelligence as well now let's talk about Finance algorithmic trading involves the use of complex artificial intelligence systems to make trading decisions at speeds several orders of magnitudes greater than any human is capable of after making millions of trades in a day without any human intervention automated trading systems are typically used by large institutional investors so here and also artificial intelligence is making things automatic now is the time to look at very fascinating applications of artificial intelligence that have been developed recently now reshma what if I tell you that the article you see on the right hand side is not written by any human it is written by words with and who is Wordsmith well Wordsmith is a natural language generation engine that lets you turn data into text at any scale and in any formatted language so all you have to do reshma is to enter new data to create a very unique story now let's look at few more applications of artificial intelligence let's talk about transport guys before that Richmond will share a few starts with us so I have a few depressing stats in front of me so nearly 1.3 million people die in Road crashes each year on an average of 3287 deaths per day and an additional 20 to 50 million are endured or disabled and more than half of all the road traffic debts occur among the young adults aged from 15 years to 44 years now you guys can imagine the wasted of human resource why do you think it happens it happens because humans tend to become negligible and care less at Times They Might Be Drug while they are driving they might get tired they might be feeling sleepy so because of all these reasons these accidents are actually increasing they're not decreasing at all so the more number of cars we have the more number of reckless drivers we have and the more number of accidents happening everyday now a possible solution to this is if a car can drive on its own now obviously a machine won't drink and drive and a machine will never get tired because of all these reasons self-driving cars are intrigues these cars make our roads safer than what it is right now now since we are talking about self-driving cars let me tell you that we even have self flying helicopters the k-max may be used for firefighting but it can carry up to 6 000 pounds of cargo this helicopter can drop nearly 2 800 gallons of water on the Fire now talking about helicopters risma has a very interesting application to share with us yes I'm very excited to tell you about this application so there is an interesting example of artificial intelligence technology that is creative for stopping hijackers so this machine combines the function of a drone a helicopter and trained snipers and this helicopter is called the vigilante and I personally love visilante so yeah bring it on Terrace we have vigilante do your worst now raise some of the painting that you're seeing is created by a very famous painter called Aaron have you heard of him well I am not a very artistic person and no I have not heard of Aaron but paintings look really amazing I think I'm a fan all right so you're a fan of Aaron which is a computer program written by artist Harold Cohen so see even machines are creative and definitely a lot creative than you serve I could sense that stress that you made in definitely yes of course and one more amazing creative application of artificial intelligence is that it can also compose music so well reshma let's play it for our viewers as well as for us definitely [Music] [Music] thank you [Music] now in 2011 we even saw IBM Watson I guess you guys have heard about IBM Watson so Watson is an IBM supercomputer that uses artificial intelligence and sophisticated analytical software for Optimal Performance as a question answering machine now reshma has very interesting stats about Watson to share with us yeah so let me tell you a few interesting details about Watson so the Watson supercomputer processes at a rate of 80 teraflops that is trillion floating Point operations per second and to replicate a high functioning human's ability to answer questions Watson accesses 90 servers with a combined data store of over 200 million pages of information which is processed against 6 million logic rules and the device and its data are self-contained in a space that could actually accommodate 10 refrigerators when you are not capable you tend to become arrogant so because of that Watson challenge the two top ranked players on Jeopardy and defeated Champions Ken Jennings and Brad Rutter in 2000 with artificial intelligence Google deepminds alphago defeated go Champion Lee sit down which artificial intelligence researcher thought was not possible to achieve in next 20 years so Raceway when I was talking about making systems that are more intelligent than us you can see that we have multiple examples like that and this game is thought to have originated at least 2 500 to 4 000 years ago and it is the oldest game that is still played in its original form I didn't know that you're welcome so now you can see that we have come a long way but trust me guys we have just finished the course of artificial intelligence we still need to explore its green and guys if you have any cool application of artificial intelligence that you want to share feel free to mention it in the comment section so now let's discuss about the artificial intelligence and deep learning with tensorflow course provided by edureka so this is how the structure of the course is guys we'll Begin by understanding deep learning we'll understand how it evolved what is the reason for it for its introduction what were we doing before that so all those things will be covered in this particular module then we'll understand how neural networks work with tensorflow now tensorflow is nothing but a python Library used for implementing deep learning models the third body will also deal with the same and by the end of whole body will be the master of deep networks after that we'll be discussing various types of networks first is convolutional neural networks that is used for image recognition then recurrent neural networks RBM and auto encoders then the eighth module will deal with Keras which is nothing but one more Library like tensorflow after that we'll be discussing high level apis like TF lower and our 10th module will be entirely dedicated to the project now you guys must be wondering what are the products so let us look at those projects one by one so the first project that you will get will be real estate price protection now in real estate price protection you'll be given a data set that will contain number of a lot of features and or you can say columns so those columns will be like you know crime rate what is the size of your house and things like that and your last column will be the label that is the house value so you need to train your model on that particular data set so whenever you feed in those inputs those features then it should predict what is the value of that particular house the next project is never mind detector so it is basically used in order to figure out whether the obstacle is a rock or is it a never mind now Neville mines are nothing but self-contained explosive devices placed in water to damage submarines and surface ships so the data set basically contains sonar signals bounced off a metal cylinder as well as sonar signals bounced off a drop so you need to train your model on that particular data set so that when you feed in an input it should determine whether it is a rock or is it a never mind so the next project will be very interesting guys it's about image recognition system so here the data set that we'll use will contain 10 types of images and multiple images of those 10 classes you need to train your model on that particular data set so that whenever you feed in an image of which belongs to those 10 classes it should be able to recognize that particular image it can be a bird can be airplane automobile and similarly there are seven other classes the next project is about digit recognition system so here we have a data set that contains handwritten digits which has 10 classes so basically we have all the digits from 0 to 9 and you need to predict or you need to train your model such that whenever we feed in an image it should recognize what digit it is similarly we have two more projects we have one which is the sentiment analysis system so you'll be training your model on an IMDb data set which will tell you whether you like a particular movie or not based on your earlier preferences and then the last project is very interesting this is a wine classifier project you'll be classifying different kinds of wine using a neural network well the voice of rational was very different when she was explaining the sixth project I think she's pretty excited about it yes I am excited about it this is a very cool project so by this we have reached the end of our sessions so if you have any questions regarding the AI and deep learning with tensorflow course you can reach out to us you can mail us you can call us you can visit our website at www.edureka.com and if you want us to reach back to you you can leave your email ID in the comment section and we'll get back to you at the earliest guys I'll leave the relevant contact details in the comment section so you can reach out to us anytime you want thank you guys and have a great day