Presentation on theme: "CS541 Artificial Intelligence"— Presentation transcript:
1 CS541 Artificial Intelligence Lecture I: Introduction and Intelligent Agent
2 华刚 Self-introduction Prof. Gang Hua Associate Professor in Computer ScienceStevens Institute of TechnologyResearch Staff Member (07/2010—08/2011)IBM T J. Watson Research CenterSenior Researcher (08/2009—07/2010)Nokia Research Center HollywoodScientist (07/2006—08/2009)Microsoft Live Labs ResearchPh.D. in ECE, Northwestern University, 06/2006
3 Course Information (1) CS541 Artificial Intelligence Term: Fall 2012 Instructor: Prof. Gang HuaClass time: Wednesday 6:15pm—8:40pmLocation: Babbio Center/Room 210Office Hour: Wednesday 4:00pm—5:00pm by appointmentOffice: Lieb/Room305Course Assistant: Yizhou LinCourse Website:cs541_artificial_intelligence_fall_2012.htm
4 Course Information (2) Text Book: Grading: Stuart Russell and Peter Norvig, “Artificial Intelligence: A Modern Approach”, Third Edition, Prentice Hall, December 11, (Required)Grading:Class Participation: 10%5 Homework: 50% (including a midterm project)Final Project & Presentation: 40%
5 Schedule Week Date Topic Reading Homework** 1 08/29/2012 Introduction & Intelligent AgentCh 1 & 2N/A209/05/2012Search: search strategy and heuristic searchCh 3 & 4sHW1 (Search)309/12/2012Search: Constraint Satisfaction & Adversarial SearchCh 4s & 5 & 6 Teaming Due409/19/2012Logic: Logic Agent & First Order LogicCh 7 & 8sHW1 due, Midterm Project (Game)509/26/2012Logic: Inference on First Order LogicCh 8s & 9610/03/2012No class 710/10/2012Uncertainty and Bayesian Network Ch 13 & Ch14s HW2 (Logic)810/17/2012Midterm PresentationMidterm Project Due910/24/2012Inference in Baysian NetworkCh 14sHW2 Due, HW3 (Probabilistic Reasoning)1010/31/2012Probabilistic Reasoning over TimeCh 151111/07/2012Machine Learning HW3 due,1211/14/2012Markov Decision ProcessCh 18 & 20HW4 (Probabilistic Reasoning Over Time) 1311/21/2012Ch 161411/29/2012Reinforcement learningCh 21HW4 due1512/05/2012Final Project PresentationFinal Project Due
6 Rules Need to be absent from class? Late submission of homework? 1 point per class: please send notification and justification at least 2 days before the classLate submission of homework?The maximum grade you can get from your late homework decreases 50% per dayZero tolerance on plagiarism!!You receive zero grade
7 Introduction & Intelligent Agent Prof. Gang HuaDepartment of Computer ScienceStevens Institute of Technology
8 Introduction to Artificial Intelligence Chapter 1
9 Systems thinking humanly Systems thinking rationally What is AI?Systems thinking humanlySystems thinking rationallySystems acting humanlySystems acting rationally
10 Acting humanly: Turing Test Turing (1950) "Computing machinery and intelligence":"Can machines think?" "Can machines behave intelligently?"Operational test for intelligent behavior: the Imitation GamePredicted that by 2000, a machine might have a 30% chance of fooling a layperson for 5 minutesAnticipated all major arguments against AI in following 50 yearsSuggested major components of AI: knowledge, reasoning, language understanding, learning,Total Turing test: adding vision and roboticsProblem: Turing test is not reproducible, constructive, or amenable to mathematical analysis
11 Thinking humanly: cognitive modeling 1960 "cognitive revolution": information-processing psychology replaced prevailing orthodoxy of behaviorismRequires scientific theories of internal activities of the brainWhat levels of abstraction? "Knowledge" or "circuits"?How to validate? RequiresPredicting and testing behavior of human subjects (top-down)Direct identification from neurological data (bottom-up)Both approaches (roughly, Cognitive Science and Cognitive Neuroscience) are now distinct from AIBoth share one principal direction with AI:The available theories do not explain anything resembling human-level general intelligence
12 Thinking rationally: "laws of thought" Aristotle: what are correct arguments/thought processes?Several Greek schools developed various forms of logic:notation and rules of derivation for thoughts;They may or may not have proceeded to the idea of mechanizationDirect line through mathematics and philosophy to modern AIProblems:Not all intelligent behavior is mediated by logical deliberationWhat is the purpose of thinking?What thoughts should I have out of all the thoughts (logical or otherwise) that I could have?
13 Acting rationally: rational agent Rational behavior: doing the right thingThe right thing: that which is expected to maximize goal achievement, given the available informationDoesn't necessarily involve thinking – e.g., blinking reflex – but thinking should be in the service of rational actionAristotle (Nicomachean Ethics):Every art and every inquiry, and similarly every action and pursuit, is thought to aim at some good
14 Rational agents An agent is an entity that perceives and acts This course is about designing rational agentsAbstractly, an agent is a function from percept histories to actions:For any given class of environments and tasks, we seek the agent (or class of agents) with the best performanceCaveat: computational limitations make perfect rationality unachievable design best program for given machine resources
15 AI prehistory Philosophy Mathematics Economics Neuroscience Psychology Logic, methods of reasoning, mind as physical system, foundations of learning, language, rationalityMathematicsFormal representation and proof algorithms, computation, (un)decidability, (in)tractability, probabilityEconomicsUtility, decision theoryNeurosciencePhysical substrate for mental activityPsychologyPhenomena of perception and motor control, experimental techniquesComputer engineeringBuilding fast computersControl theoryDesign systems that maximize an objective function over timeLinguisticsknowledge representation, grammar
16 Abridged history of AIMcCulloch & Pitts: Boolean circuit model of brainTuring's "Computing Machinery and Intelligence"1956 Dartmouth meeting: "Artificial Intelligence" adopted1952—69 Look, Ma, no hands!1950s Early AI programs, including Samuel's checkers program, Newell & Simon's Logic Theorist, Gelernter's Geometry Engine1965 Robinson's complete algorithm for logical reasoning1966—74 AI discovers computational complexity Neural network research almost disappears1969—79 Early development of knowledge-based systems1980—88 Expert systems industry boom1988—93 Expert systems industry busts: "AI Winter"1985—95 Neural networks return to popularity1988— Resurgence of probability; AI becomes science1995— The emergence of intelligent agents2003— Human-level AI back on the agenda
17 State of the artDeep Blue defeated the reigning world chess champion Garry Kasparov in 1997Proved a mathematical conjecture (Robbins conjecture) unsolved for decadesNo hands across America (driving autonomously 98% of the time from Pittsburgh to San Diego)During the 1991 Gulf War, US forces deployed an AI logistics planning and scheduling program that involved up to 50,000 vehicles, cargo, and peopleNASA's on-board autonomous planning program controlled the scheduling of operations for a spacecraftProverb solves crossword puzzles better than most humansiRobot corporated in 2000: Roomba & ScoobaGoogle cars automatically are driving in the city to collect stree-tview imagesWatson whips Brad Rutter and Ken Jennings in Jeopardy in 2011!
20 Outline Agents and environments Rationality: what is a rational agent? PEAS (Performance measure, Environment, Actuators, Sensors)Environment typesAgent types
21 AgentsAn agent is anything that can be viewed as perceiving its environment through sensors and acting upon that environment through actuatorsHuman agent: eyes, ears, and other organs for sensors; hands, legs, mouth, and other body parts for actuatorsRobotic agent: cameras and infrared range finders for sensors; various motors for actuators
22 Agents and environments The agent function maps from percept histories to actions:The agent program runs on the physical architecture to produce fagent = architecture + program
23 Vacuum-cleaner world Percepts: location and contents, e.g., [A,Dirty] Actions: Left, Right, Suck, NoOp
24 Rational agents (1)An agent should strive to "do the right thing", based on what it can perceive and the actions it can perform. The right action is the one that will cause the agent to be most successfulPerformance measure: An objective criterion for success of an agent's behaviorE.g., performance measure of a vacuum-cleaner agent could be:Amount of dirt cleaned up in time T?Amount of dirt cleaned up minus the amount of electricity consumed in time T?Amount of time taken to clean a fixed region?amount of dirt cleaned up, amount of time taken, amount of electricity consumed, amount of noise generated, etc.
25 Rational agents (2)Rational Agent: For each possible percept sequence, a rational agent should select an action that is expected to maximize its performance measure, given the evidence provided by the percept sequence and whatever built-in knowledge the agent has.
26 Rational agents (3)Rationality is distinct from omniscience (all-knowing with infinite knowledge)Agents can perform actions in order to modify future percepts so as to obtain useful information (information gathering, exploration)An agent is autonomous if its behavior is determined by its own experience (with ability to learn and adapt)
27 PEAS (1) PEAS: Performance measure, Environment, Actuators, Sensors To design a rational agent, we must first specify the task environmentConsider, e.g., the task of designing an automated taxi driver:Environment??Actuators??Sensors??
28 PEAS (2)To design a rational agent, we must first specify the task environmentConsider, e.g., the task of designing an automated taxi driver:Sensors: Cameras, sonar, speedometer, GPS, odometer, engine sensors, keyboard
29 PEAS (3) Agent: Internet shopping agent Performance measure: price, quality, appropriateness, efficiencyEnvironment: current and future WWW sites, vendors, shippersActuators: display to user, follow URL, fill in formSensors: HTML pages (text, graphics, scripts)
30 PEAS (4) Agent: Part-picking robot Performance measure: Percentage of parts in correct binsEnvironment: Conveyor belt with parts, binsActuators: Jointed arm and handSensors: Camera, joint angle sensors
31 PEAS (5) Agent: Interactive English tutor Performance measure: Maximize student's score on testEnvironment: Set of studentsActuators: Screen display (exercises, suggestions, corrections)Sensors: Keyboard
32 Environment types (1)Fully observable (vs. partially observable): An agent's sensors give it access to the complete state of the environment at each point in time.Deterministic (vs. stochastic): The next state of the environment is completely determined by the current state and the action executed by the agent. (If the environment is deterministic except for the actions of other agents, then the environment is strategic)Episodic (vs. sequential): The agent's experience is divided into atomic "episodes" (each episode consists of the agent perceiving and then performing a single action), and the choice of action in each episode depends only on the episode itself.
33 Environment types (2)Static (vs. dynamic): The environment is unchanged while an agent is deliberating. (The environment is semidynamic if the environment itself does not change with the passage of time but the agent's performance score does)Discrete (vs. continuous): A limited number of distinct, clearly defined percepts and actions.Single agent (vs. multiagent): An agent operating by itself in an environment.
34 Environment types (3)SolitaireBackgammonInternet ShoppingTaxiObservable?YesNoDeterministic?PartlyEpisodic?Static?SemiDiscrete?Single Agent?Yes (except auction)The environment type largely determines the agent designThe real world is (of course) partially observable, stochastic, sequential, dynamic, continuous, multi-agent
35 Agent functions and programs An agent is completely specified by the agent function mapping percept sequences to actionsAim: find a way to implement the rational agent function concisely
36 Table-lookup agent Drawbacks: Huge table Take a long time to build the tableNo autonomyEven with learning, need a long time to learn the table
37 A vacuum-cleaner agent What is the right function?Can it be implemented in a small agent program?
38 Agent types Four basic types (with increasing generality): Simple reflex agentsModel-based reflex agentsGoal-based agentsUtility-based agentsAll of them can be transformed into learning agent
39 Simple reflex agentsThe action to be selected only depends on the most recent percept, not a sequenceThese agents are stateless devices which do not have memory of past world states
40 Model-based reflex agents Have internal state which is used to keep track of past states of the worldCan assist an agent deal with some of the unobserved aspects of the current state
41 Goal-based agentsAgent can act differently depending on what the final state should look likeE.g., automated taxi driver will act differently depending on where the passenger wants to go
42 Utility-based agentsAn agent's utility function is an internalization of the external performance measureThey may differ if the environment is not completely observable or deterministic
43 Learning agentsLearning agent cuts across all of the other types of agents: any kind of agent can learn
45 SummaryAgents interact with environments through actuators and sensorsThe agent function describes what the agent does in all circumstancesThe performance measure evaluates the environment sequenceA perfectly rational agent maximizes expected performanceAgent programs implement (some) agent functionsPEAS descriptions define task environmentsEnvironments are categorized along several dimensions:Observable? Deterministic? Episodic? Static? Discrete? Single-agent?Several basic agent architectures exist:Reflex, Reflex with state, goal-based, utility-based