Re Repr pres esent entati ationa onal l Di Dime mens nsio - PowerPoint PPT Presentation

Re Repr pres esent entati ationa onal l Di Dime mens nsio ions ns Computer Co ter Sc Science ce cpsc3 c322, 22, Lectur ture e 2 (Te Text xtbo book ok Chpt1) January, ary, 6, 2010 CPSC 322, Lecture 2 Slide 1

Depa partme ment nt of Compu puter ter Science nce Undergr dergrad aduat uate e Events ts Drop-In In Resume Edition Session Events ts this week Date: : Mon. Jan 11 How to Prepare for the Tech Career Time: : 11 am – 2 p pm Fair Location ion: : Rm Rm 255, ICICS/C /CS S Bldg Date: : Wed. Jan 6 Time: : 5 – 6:30 pm Industry stry Panel Location ion: : DMP 110 Speakers: rs: Managers rs from Google, , Resume Writing Workshop (for non- IBM, , Microso soft ft, , TELUS, , etc. coop students nts) Date: : Tues. Jan 12 Date: : Thurs. . Jan 7 Time: : Panel: 5:15 – 6:15 pm; Time: : 12:30 – 2 p pm Networki rking ng: : 6:15 – 7:15 pm Location ion: : DMP 201 Location ion: : Panel: DMP 110; Networki rking ng: : X-wing Undergrad CSSS Movie Night Lounge Lounge Date: : Thurs. s. Jan 7 Time: : 6 – 10 pm Tech Career Fair Location ion: : DMP 310 Date: : Wed. Jan 13 Movies: “Up” & “The Hangover” Time: : 10 am – 4 p pm (Free e Popcorn & Pop) Location ion: : SUB Ballroom

Lecture cture Ov Overview view • Rec ecap ap fr from om la last t le lectu ture re • Representation and Reasoning • An Overview of This Course • Further Dimensions of Representational Complexity CPSC 322, Lecture 2 Slide 3

Course urse Essential entials • Course se web-pag age e : CHECK IT OFTEN! • Te Textbo tbook ok: Available online and pdf on WebCT • We will cover at least Chapters: 1, 3, 4, 5, 6, 8, 9 • WebCT: T: used for textbook, discussion board…. • AI AIspac ace e : online tools for learning Artificial Intelligence http://aispace.org/ • Lecture slides… • Midter erm exam, Wed, Mar 10 (1.5 hours, regular room) Any conflict? CPSC 322, Lecture 2 Slide 4

Agents ents acti ting ng in an environ ironment ment Representation & Reasoning CPSC 322, Lecture 2 Slide 5

Lecture cture Ov Overview view • Recap from last lecture • Rep epresen esentati tation on an and Re d Reas ason onin ing • An Overview of This Course • Further Dimensions of Representational Complexity CPSC 322, Lecture 2 Slide 6

Wh What t do we need ed to to represent resent ? • Th The enviro ronm nmen ent t /world d : What different configurations (states tes / possi sible le worlds) can the world be in, and how do we denote them? Chessboard, Info about a patient, Robot Location • How the world works s (we we wi will focus s on) • Co Constr strain aints: ts: sum of current into a node = 0 • Causal al: : what are the causes and the effects of brain disorders? • Ac Actions ons preconditions and effects: when can I press this button? What happens if I press it? CPSC 322, Lecture 2 Slide 7

Corresponding rresponding Reasoning soning Ta Tasks s / / Pr Problems oblems • Const strai aint nt Sa Satisfactio faction – Fi Find state te that satis isfie fies s set of constra train ints ts. E.g., What is a feasible schedule for final exams? • An Answe werin ring Qu Query – Is a given propositi sition on true/l e/like ikely ly given en what is known? ? E.g., Does this patient suffers from viral hepatitis? • Pl Plannin ing g – Fi Find sequence nce of actio ions ns to reach a goal state te / maximize ze utility ity. E.g., Navigate through and environment to reach a particular location CPSC 322, Lecture 2 Slide 8

Representation presentation and Reasoning soning System tem • A (repres esentatio ntation) language ge in which the environment and how it works can be described • Computational (reason onin ing) proced edur ures es to compute a solution to a problem in that environment (an answer, a sequence of actions) But the choice of an appropriate R&R system Bu depends on a key property of the environment and of the agent’s knowledge CPSC 322, Lecture 2 Slide 9

Deterministic terministic vs. . Sto tochastic hastic (Uncertain) certain) Domains mains • Se Sensin ing g Uncertai tainty nty: Can the agent fully observe the current state of the world? • Ef Effec ect t Uncertai tainty nty: Does the agent knows for sure what the effects of its actions are? Doctor Diagnosis Poker Factory Floor Chess Doctor Treatment CPSC 322, Lecture 2 Slide 10

Deterministic terministic vs. . Sto tochastic hastic Domains ains Historically, AI has been divided into two camps: those who prefer representations based on logic and those who prefer probability ility. A few years ago, CPSC 322 covered logic, while CPSC 422 introduced probability: • now we introduce both representational families in 322, and 422 goes into more depth • this should give you a better idea of what's included in AI Note: Some of the most exciting current research in No AI is actually building bridges between these camps. CPSC 322, Lecture 2 Slide 11

Lecture cture Ov Overview view • Recap from last lecture • Representation and Reasoning • An n Overv erview iew of of Thi his Cou ourse se • Further Dimensions of Representational Complexity CPSC 322, Lecture 2 Slide 12

Modules dules we'l 'll l cover er in th this course: se: R&Rsys sys Enviro En ronm nmen ent Stochastic Deterministic Pr Problem em Arc Consistency Constraint Vars + Satisfaction Search Constraints Stati atic Belief Nets Logics Query Var. Elimination Search Decision Nets Sequenti ntial al STRIPS Var. Elimination Planning Markov Processes Search Representation Value Iteration Reasoning CPSC 322, Lecture 2 Slide 13 Technique

Lecture cture Ov Overview view • Recap from last lecture • Representation • An Overview of This Course • Fu Furth ther er Dim imen ensions ions of of Rep epres esentatio entationa nal l Com ompl plexity exity CPSC 322, Lecture 2 Slide 14

Dim imensi ensions ons of of Rep epres esen entation tational al Com ompl plex exity ity We'v 've already dy discu cuss ssed: ed: • Reasoning tasks (Static vs. Sequential ) • Deterministic versus stochastic domains So Some other r importan tant t dimensi sion ons s of complex exity: ity: • Explicit state or propositions or relations • Flat or hierarchical • Knowledge given versus knowledge learned from experience • Goals versus complex preferences • Single-agent vs. multi-agent CPSC 322, Lecture 2 Slide 15

Expli plicit cit Sta tate te or propositions positions How do we model the environment? • You can enumerate the states tes of the world. • A state can be described in terms of featur tures es • Often it is more natural to describe states in terms of assignments of values to features (variables). • 30 binary features (also called propositions) can represent 2 30 = 1,073,741,824 states. Mars Ex Explorer er Ex Example Weather Temperature LocX LocY CPSC 322, Lecture 2 Slide 16

Expli plicit cit Sta tate te or propositions positions or relations ations • States can be described in terms of objects cts and relati ation onsh ship ips. • There is a proposition for each relationship on each “possible” tuple of individuals. Univer ersi sity y Ex Example Registred(S,C) Students (S) = { } Courses (C) = { } • Textbook example: One binary relation and 10 individuals can represents 10 2 =100 propositions and 2 100 states! CPSC 322, Lecture 2 Slide 17

Fl Flat t or hierarchical rarchical Is it useful to model the whole world at the same level of abstraction? • You can model the world at one level of abstraction: flat at • You can model the world at multiple levels of abstraction: hierarc rchi hica cal • Example: Planning a trip from here to a resort in Cancun, Mexico CPSC 322, Lecture 2 Slide 18

Kn Know owle ledg dge e gi given en vs. . kno nowle ledg dge e le lear arne ned d fr from om ex expe perience ience The agent is provided with a model of the world once and far all • The agent can learn how the world works based on experience • in this case, the agent often still does start out with some prior knowl wledge CPSC 322, Lecture 2 Slide 19

Go Goals als versus sus (complex) plex) prefer ferences ences An agent may have a goal goal that it wants to achieve • e.g., there is some state te or set of states tes of the world that the agent wants to be in • e.g., there is some proposition ition or set of propositi sition ons that the agent wants to make true An agent may have prefer eren ence ces • e.g., there is some preferen erence/ut ce/utility ity functi ction that describes how happy the agent is in each state of the world; the agent's task is to reach a state which makes it as happy as possible Preferences can be complex… What beverage to order? • The sooner I get one the better • Cappuccino better than Espresso CPSC 322, Lecture 2 Slide 20