Introduction to Artificial Intelligence COMP 30030 Dr Robin Burke Who am I Associate Professor at DePaul University Chicago Illinois Fulbright Scholar at UCD for 2008 2009 Academic Year PhD from Northwestern University Evanston Illinois 1993 Artificial Intelligence Research Case based reasoning Recommender systems User modeling Teaching Mostly computer game development including game AI About this class Marking Exam 50 Mid term 15 Practicals 35 Project Homework 20 15 Lectures Tuesday Thursday 9 10 Computer Science Theatre Bear with me First time teaching here Many things I don t know we will rely a great deal on our teaching assistant Contact information rburke cs depaul edu most reliable Office phone 716 2858 TA contact info Not assigned yet Moodle forum will also be important Book Russell Norvig Artificial Intelligence A Modern Approach cover chapters 1 6 skipping some bits Other books probably cover the core topics as well What is AI problematic concept conjunction of two difficult terms artificial not natural intelligence Typical definition Minsky The science of making machines do things that would require intelligence if done by people Problems with the definition Cube roots if you can do cube roots in your head most people would say that requires intelligence no one would call numeric algorithms AI Vision chickens can recognize complex objects in visual scenes duplicating such abilities is a not so easy AI problem we don t think chickens are intelligent Optical character recognition used to be considered an AI problem then it was mostly solved now we don t consider it AI My favorite definition courtesy Chris Riesbeck Artificial Intelligence is the search for the answer to the fundamental question Why are computers so stupid Computer stupidity Examples crash performance problems anomalies not noticed Some AI topics Learning how to get computers to do better over time and experience how to get them to take correction instruction Common sense how to get computers to know basic facts about the world can a ping pong ball be used as a football can you drive from Dublin to Chicago Language how to get computers to make use of human language The man with the white dog with the white gloves The man with the white dog with the white tail Mike thinks chocolate Planning how to get computers to act in the world starting from Celbridge get to UCD in time for class AI philosophies Hard Strong AI The result being sought is a computer program that exhibits intelligent behavior and the program constitutes an explanation of how humans achieve the same behavior Example a computer vision algorithm might claim to reproduce certain functions in the human visual system Weak AI We re interested in solving the problem but we don t care how people do it How could a program explain human intelligence Obviously people and computers are very different neurons vs transistors sugars vs electric current neurotransmitter diffusion vs voltage pulses AI programs will not be the same as brains so is hard AI doomed Not necessarily some would disagree birds fly and so do airplanes they don t fly the same way on the other hand airplanes solve the problem of moving people through the air and have taught us about the fundamental principles that help us understand bird flight functional equivalence Some related fields Cognitive science branch of psychology interested in modeling intelligence controlled experiments focused on basic features of thought example 7 2 Names Pinker Tversky Linguistics study of language commonalities across languages suggest relationships between language and thought example shape words are learned earlier than color words Names Chomsky Lakoff Philosophy What is knowledge How do we know How do we know that we know What is thought intelligence Names Dennett Dreyfus The Turing Test Proposed in 1950 by Alan Turing Given only a textual interface Can a experimental subject tell the difference between a computer and a person Conversation could be about anything What are you wearing today Is Ronaldinho over the hill Estimated that by 2000 a computer would have 30 chance of fooling a lay person for 5 minutes not quite there Raises many research issues that are still open questions Some dislike it the computer has to lie computer has to match human cognitive limitations ELIZA Written in the 1960 s by Joseph Weizenbaum Engages in conversations with a user example Eliza s responses are generated through pattern matching Eliza s memory contains a set of patterns and responses Transformed input is matched against the patterns whichever pattern matches determines the response Occasionally inputs are stored so that they can be responded to later The experience made Weizenbaum a skeptic of the Turing test it obviously wasn t intelligent but some people did respond to it as if it were human Loebner Prize Annual competition similar to the Turing test Limitations placed on questioners make the results somewhat suspect 2007 winner UltaHal AI Research No one works on general purpose Turing intelligence a Turing capable system would need to integrate all of the AI capabilities learning language planning etc all still open problems Cyc comes closest AI research is specialized focused on specific problems focused on adding specific capabilities to systems often the AI component is a small but crucial subpart planning on NASA missions Many AI successes Everyday AI Spell grammar checkers Medical diagnosis systems Regulating Controlling hardware devices and processes e g in automobiles Voice image recognition more generally pattern recognition Scheduling systems airlines hotels manufacturing Program verification compiler and programming language design Web search engines Web spiders Web personalization Recommender systems collaborative content filtering Credit card verification in e commerce fraud detection Data mining and knowledge discovery in databases Computer games Computer Games In games the non player characters are often referred to as AI The AIs will retreat if you throw grenades Game AI is quite different from real AI Game AI often covers techniques that are not considered AI like AI uses techniques impractical in a game context AI Game AI Differences Analogy game AI is to real AI as stage design is to architecture The goal of game AI is to give the impression of intelligence to avoid the impression of stupidity to provide a reasonable challenge for the player Challenge It is very possible to make the