Unit 1w, 3 Applications of Ai

So the correct answer is all of those-- finance, robotics, games, medicine, the Web, and many more applications. So let me talk about them in some detail. There is a huge number of applications of artificial intelligence in finance, very often in the shape of making trading decisions-- in which case, the agent is called a trading agent. And the environment might be things like the stock market or the bond market or the commodities market. And our trading agent can sense the course of certain things, like the stock or bonds or commodities. It can also read the news online and follow certain events. And its decisions are usually things like buy or sell decisions--trades. There's a huge history of artificial intelligence finding methods to look at data over time and make predictions as to how courses develop over time-- and then put in trades behind those. And very frequently, people using artificial intelligence trading agents have made a good amount of money with superior trading decisions. There's also a long history of AI in Robotics. Here is my depiction of a robot. Of course, there are many different types of robots and they all interact with their environments through their sensors, which include things like cameras, microphones, tactile sensor or touch. And the way they impact their environments is to move motors around, in particular, their wheels, their legs, their arms, their grippers. They can also say things to people using voice. Now there's a huge history of using artificial intelligence in robotics. Pretty much, every robot that does something interesting today uses AI. In fact, often AI has been studied together with robotics, as one discipline. But because robots are somewhat special in that they use physical actuators and deal with physical environments, they are a little bit different from just artificial intelligence, as a whole. When the Web came out, the early Web crawlers were called robots and to block a robot from accessing your website, to the present day, there's a file called robot.txt, that allows you to deny any Web crawler to access and retrieve that information from your website. So historically, robotics played a huge role in artificial intelligence and a good chunk of this class will be focusing on robotics. AI has a huge history in games-- to make games smarter or feel more natural. There are 2 ways in which AI has been used in games, as a game agent. One is to play against you, as a human user. So for example, if you play the game of Chess, then you are the environment to the game agent. The game agent gets to observe your moves, and it generates its own moves with the purpose of defeating you in Chess. So most adversarial games, where you play against an opponent and the opponent is a computer program, the game agent is built to play against you--against your own interests--and make you lose. And of course, your objective is to win. That's an AI games-type situation. The second thing is that games agents in AI also are used to make games feel more natural. So very often games have characters inside, and these characters act in some way. And it's important for you, as the player, to feel that these characters are believable. There's an entire sub-field of artificial intelligence to use AI to make characters in a game more believable--look smarter, so to speak-- so that you, as a player, think you're playing a better game. Artificial intelligence has a long history in medicine as well. The classic example is that of a diagnostic agent. So here you are--and you might be sick, and you go to your doctor. And your doctor wishes to understand what the reason for your symptoms and your sickness is. The diagnostic agent will observe you through various measurements-- for example, blood pressure and heart signals, and so on-- and it'll come up with the hypothesis as to what you might be suffering from. But rather than intervene directly, in most cases the diagnostic of your disease is communicated to the doctor, who then takes on the intervention. This is called a diagnostic agent. There are many other versions of AI in medicine. AI is used in intensive care to understand whether there are situations that need immediate attention. It's been used for life-long medicine to monitor signs over long periods of time. And as medicine becomes more personal, the role of artificial intelligence will definitely increase. We already mentioned AI on the Web. The most generic version of AI is to crawl the Web and understand the Web, and assist you in answering questions. So when you have this search box over here and it says "Search" on the left, and "I'm Feeling Lucky" on the right, and you type in the words, what AI does for you is it understands what words you typed in and finds the most relevant pages. That is really co-artificial intelligence. It's used by a number of companies, such as Microsoft and Google and Amazon, Yahoo, and many others. And the way this works is that there's a crawling agent that can go to the World Wide Web and retrieve pages, through just a computer program. It then sorts these pages into a big database inside the crawler and also analyzes developments of each page to any possible query. When you then come and issue a query, the AI system is able to give you a response-- for example, a collection of 10 best Web links. In short, every time you try to write a piece of software, that makes your computer software smart likely you will need artificial intelligence. And in this class, Peter and I will teach you many of the basic tricks of the trade to make your software really smart.