Tucker Hermans. Introduction to AI. CS 6300 Artificial Intelligence Spring 2018 Tucker Hermans

Transcription

1 Introduction to AI CS 6300 Artificial Intelligence Spring 2018 Tucker Hermans

2 Today What is AI? What can AI do? What is this course?

3 : Early Days of AI 1943: McCulloch & Pitts: Boolean circuit model of brain (artificial neurons) 1949: Hebb developed a rule for modifying connection strengths between neurons. 1951: Minsky and Edmonds built the first neural network computer vacuum tubes and a surplus automatic pilot mechanisms from a B-24 bomber to simulate a network of 40 neurons! 1950: Turing Computing machinery and intelligence

4 1950: Turing Test Can machines think? Can machines behave intelligently? Operational test for intelligent behavior: the Imitation Game Predicted by 2000, a 30% chance of fooling a lay person for 5 minutes Anticipated all major arguments against AI in following 50 years Suggested major components of AI: knowledge, reasoning, language understanding, learning. Problem: Turing test is not reproducible or amenable to mathematical analysis.

5 : Look, ma, no hands! Period Computers were not very powerful, so it was exciting when they did anything remotely intelligent. 1950s: Early AI programs, including Samuel's checkers program Newell & Simon's Logic Theorist Gelernter's Geometry Engine 1954 Devol and Engelberger design the first programmable robot arm, start Unimation 1956: Dartmouth meeting: Artificial Intelligence adopted 1958: McCarthy developed LISP, the second-oldest programming language in current use (FORTRAN is one year older) 1965: Robinson's complete algorithm for logical reasoning E.g., generate plan for driving to the airport 1966: Weizenbaum's Eliza / Turing test

6 The Dartmouth Conference (1956) The term artificial intelligence was coined at a gathering of researchers at Dartmouth College: We propose that a two-month, ten-man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. Many of the founders of AI were in attendance: John McCarthy, Marvin Minsky, Allen Newell, Arthur Samual, Herb Simon.

7 Herb Simon (1957) It is not my aim to surprise or shock you but the simplest way I can summarize is to say that there are now in the world machines that think, that learn, and that create. Moreover, their ability to do these things is going to increase rapidly into a visible future the range of problems they can handle will be coextensive with the range to which human mind has been applied. More precisely: within 10 years a computer would be chess champion, and an important new mathematical theorem would be proved by a computer.

8 Harder than originally thought Herb Simon's predictions came true, but after ~40 years instead of 10. Eliza mother... Tell me more about your family I wanted to adopt a puppy, but it's too young to be separated from its mother??? 1957: Sputnik Automatic Russian English translation Famous example: The spirit is willing but the flesh is weak. E R E: The vodka is strong but the meat is rotten.

9 Observations Need some understanding about the world Computational tractability, NP-completeness, exponential scaling.

10 A (Short) History of AI (cont) : Knowledge-based approaches : Early development of knowledge-based systems 1970 SRI Shakey: First Intelligent Robot : Expert systems industry booms : Expert systems industry busts: AI Winter

11 Embodiment Behavior Based Robotics Rod Brooks Elephants Don t Play Chess 1988 In this paper we argue that the symbol system hypothesis upon which classical AI is base is fundamentally flawed Nouvelle AI is based on the physical grounding hypothesis. This hypothesis states that to build a system that is intelligent it is necessary to have its representations grounded in the physical world. The new methodology bases its decomposition of intelligence into individual behavior generating modules, whose coexistence and co-operation let more complex behaviors emerge.

12 A (Short) History of AI (cont) : Knowledge-based approaches : Early development of knowledge-based systems : Expert systems industry booms : Expert systems industry busts: AI Winter 1988 : Statistical approaches Resurgence of probability, focus on uncertainty General increase in technical depth Agents and learning systems AI Spring? : Where are we now?

13 What Can AI Do Today? Quiz: Which of the following can be done at present? Play a decent game of table tennis? Drive safely along a curving mountain road? Drive safely up to Alta in a snowstorm? Buy a week's worth of groceries on the web? Buy a week's worth of groceries at Smith's? Discover and prove a new mathematical theorem? Converse successfully with another person for an hour? Perform a complex surgical operation? Unload a dishwasher and put everything away? Translate spoken Chinese into spoken English in real time? Write an intentionally funny story?

14 Unintentionally Funny Stories One day Joe Bear was hungry. He asked his friend Irving Bird where some honey was. Irving told him there was a beehive in the oak tree. Joe walked to the oak tree. He ate the beehive. The End. Henry Squirrel was thirsty. He walked over to the river bank where his good friend Bill Bird was sitting. Henry slipped and fell in the river. Gravity drowned. The End. Once upon a time there was a dishonest fox and a vain crow. One day the crow was sitting in his tree, holding a piece of cheese in his mouth. He noticed that he was holding the piece of cheese. He became hungry, and swallowed the cheese. The fox walked over to the crow. The End. [Shank, Tale-Spin System, 1984]

15 Intentionally Funny Jokes Petrovic and Matthews 2013 Unsupervised joke generation form big data I like my X like I like my Y, Z. Examples: I like my relationships like I like my source, open I like my coffee like I like my war, cold I like my boys like I like my sectors, bad Human Jokes funny 33% of time Computer jokes funny 16% of time

16 Natural Language Speech technologies Automatic speech recognition (ASR) Text-to-speech synthesis (TTS) Dialog systems Language processing technologies Machine translation: Aux dires de son président, la commission serait en mesure de le faire According to the president, the commission would be able to do so. Il faut du sang dans les veines et du cran. We must blood in the veines and the courage. Information extraction Information retrieval, question answering Text classification, spam filtering, etc

17 Vision There is a person / There is an airplane The person is standing on grass The person is standing in a park The airplane is not flying Is the airplane broken? Is that Suzanne? Image courtesy Kristen Graumann

18 Robotics Robotics Part mech. eng. Part AI Reality much harder than simulations! Technologies Vehicles Rescue Soccer! Lots of automation Images from stanfordracing.org, CMU RoboCup, Honda ASIMO sites

19 Robot Videos Reinforcement learning Stefan Schaal ( Bayes nets Imitation-learning pole balancing [video] Russ Tedrake ( Perching glider [video] Andrew Ng ( Autonomous helicopter flight [video] Paul Newman ( Experience-based navigation [video]

20 Logic Logical systems Theorem provers NASA fault diagnosis Question answering Methods: Deduction systems Constraint satisfaction Satisfiability solvers (huge advances here!) Image from Bart Selman

21 Game Playing May, '97: Deep Blue vs. Kasparov First match won against world-champion Intelligent creative play 200 million board positions per second! Humans understood 99.9 of Deep Blue's moves Can do about the same now with a big PC cluster Open question: How does human cognition deal with the search space explosion of chess? Or: how can humans compete with computers at all?? 1996: Kasparov Beats Deep Blue I could feel --- I could smell --- a new kind of intelligence across the table. 1997: Deep Blue Beats Kasparov Deep Blue hasn't proven anything. Text from Bart Selman, image from IBM s Deep Blue pages

22 Game Playing Now 2007: Checkers is solved 2016: AlphaGo defeats Lee Sedol in Go

23 A Note on Deep Learning Deep neural networks in many forms are becoming popular Main advantage they allow the computer to learn the correct features / representation for the problem Really orthogonal to this course, can use them in certain places, but we wont be covering them in any detail

24 Decision Making Many applications of AI: decision making Scheduling, e.g. airline routing, military Route planning, e.g. google maps Medical diagnosis Automated help desks Fraud detection Spam classifiers Web search engines Movie and book recommendations the list goes on.

25 What is AI? The science of making machines that: Think like humans Think rationally Act like humans Act rationally

26 Rational Decisions Where does the word rational come from? from ratio (genitive rationis) "reckoning, calculation, reason" We'll use the term rational in a particular way: Rational: maximally achieving pre-defined goals Rational only concerns what decisions are made (not the thought process behind them) Goals are expressed in terms of the utility of outcomes Being rational means maximizing your expected utility Computational rationality maybe a better title for this course

27 Maximize Your Expected Utility Tucker Hermans

28 Designing Rational Agents An agent is an entity that perceives and acts. A rational agent selects actions that maximize its utility function. Characteristics of the percepts, environment, and action space dictate techniques for selecting rational actions. This course is about: General AI techniques for a variety of problem types Learning to recognize when and how a new problem can be solved with an existing technique

29 Course Topics Part I: Making Decisions Fast search Predicate Logic Logic based planning Constraint satisfaction Adversarial and uncertain search (Game Playing) Part II: Sequential Decision Making with Uncertainty Markov Decision Processes Reinforcement learning Hidden Markov Models (HMMs) Part III: Reasoning Under Uncertainty Probabilistic Graphical Models Bayes Nets Factor Graphs Decision theory Partially Observable Markov Decision Processes (POMDPs)

30 Course Mechanics More details on the syllabus Project 0 do Friday very simple python intro Programming Projects (4) Can work in groups of 2 Still need to upload your own solutions Problem Set Homeworks (~8) Every other week or so Short problem sets to reinforce topics in class Upload typed PDFs Exams 2 during semester Final on day set by registrar s office Labs A few days of class will be used for working on projects Tucker Hermans