2.ModelBased Reflex Agent Intro

Embed Size (px)

Citation preview

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    1/18

    COMP 4640

    Intelligent & Interactive Systems

    Programs Supporting

    Model - Based Reflex Agents

    November 2008

    Dr. Cheryl Seals

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    2/18

    2

    Simple reflex agents

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    3/18

    Programs that support

    Model - based Reflex Agents

    Simple reflex agents

    select precepts based on the

    current perceptignoring the rest of the precepthistory

    Example: Beetle

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    4/18

    4

    Model-based reflex agents

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    5/18

    Programs that support

    Model - based Reflex Agents

    Most systems are based on condition-action rules

    (i.e. situation-action rules, productions, or if-then rules)

    (e.g. If car-in-front is braking then initiate-braking p46)

    Model-Based Reflex Agents

    Most effective way to keep track of the part of the

    world it cant see now.

    Maintain some internal state that depends onpercept history and thereby reflects at least someof the unobserved aspects of the current state (e.g.

    using some type of variable).

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    6/18

    Production Based SystemsThe production rule paradigm originated in the field of AI with theexpert systems rule languages such as OPS5 (Brownston et al. 1985) condition action

    An inference engine cycles through all the rules in the systemmatching the condition parts of the rules with data in working memory.

    Of all the rules that match (the candidate set), one is selected usingsome conflict resolution policy and this selected rule is fired, that is,its action part is executed.

    The action part may modify the working memory, possibly accordingto the matched data and the cycle continues until no more rulesmatch.

    Rule based Rules have special ops:

    Fire, which causes a rule to be triggeredEnable, which causes a rule to be activatedDisable, which causes a rule to be deactivated

    Conflict resolution Break ties with Specification, Sequencing, Meta rules

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    7/18

    Production Based SystemsCLIPS

    (CLanguage Integration Production System) Production system developed at NASAs Johnson space

    center. Written in ANSI C instead of LISP CLIPS implements standard forward-chaining pattern-

    matching algorithm CLIPS knowledge representation similar to OPS5 and ART

    systems. Constructs

    simple string fact assertion & retractionTemplatesIf-then rules (productions)Objects and instances

    NASA uses clips in the following projects

    Intelligent computer aided crew training, weatherforecasting, shuttle space planning, shuttlediagnostics, Mission Control Center (telemetry dataanalysis and diagnostics), flight assistance andcontrol

    ART commercial expert system has many of thesame features as CLIPS

    http://portal.acm.org/author_page.cfm?id=81100007064&coll=GUIDE&dl=GUIDE&trk=0&CFID=9450350&CFTOKEN=41813484http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=00176973http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=00176973http://portal.acm.org/author_page.cfm?id=81100007064&coll=GUIDE&dl=GUIDE&trk=0&CFID=9450350&CFTOKEN=41813484
  • 8/3/2019 2.ModelBased Reflex Agent Intro

    8/18

    Agent Based Systems

    Systems to investigate

    Stagecast CreatorTM (www.stagecast.com)

    AgentsheetsTM (www.agentsheets.com)

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    9/18

    End User Programming with agentsStagecast Study Report:

    We are attempting to create a cross-generational webbased learning community for middle school students,teachers, and seniors.

    Learning community will design, construct, and discusssimulations of community issues.

    Summary of results of formative evaluation with studentscreating simulation projects.

    Proceedings of IEEE Visual Languages 2001, Rosson, Seals2001; CHI 2001; DIS 2002; NSF Research: NSF ITR 0091102.

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    10/18

    Based on a movie metaphorProgramming is facilitated by macro recorder to allowprogramming by demonstration

    Behaviors are represented as a set of as a set of

    productions or if-then rules

    Stagecast Creator

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    11/18

    Procedure

    Participants: 10 middle school students

    Background survey

    Performed in usability testing lab study with think aloud

    protocol

    Recorded critical incidents

    Captured video, audio, and screen

    Subjective questionnaire, knowledge survey,retrospective interview

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    12/18

    Visual Agent Programming

    Spatial context and visual appearance arerequired elements in a rules precondition

    Correct position and appearance are

    preconditions for rules

    If Precondition is satisfied, Then rule is fired.

    Characters may have many instantiations

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    13/18

    Observations and Results

    Duration 30-55 minutes Activity I

    Duration 34-47 minutes Activity II

    Most students were successful in modifying simulationsand adding new characters.

    Usability satisfaction

    Easy and fun to use

    Would like to use it in their classes But needed more exposure to feel confident

    No problems with drawing tools

    Problems with tools for rule creation

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    14/18

    Issue Likely Cause

    Directing input to the wrongwindow Too many similar-lookingwindows

    Confused between rules andrule-actions lists

    Lists that look similar but havedifferent meanings

    Select wrong icon Multiple similar iconsInability to find rules or othercontent in window

    Non-traditional method ofscrolling

    Misunderstand spotlight andconcept of stretching it

    Spotlight metaphor is notobvious or intuitive

    Stagecast Usability Problems

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    15/18

    Practical metaphors for icons

    Bigger Icons

    Fewer layers of scaffoldingRelation between internal variables andvisual state of the simulation.

    Role of visual context in rules Rules must match exact visual context, most

    PBD system make rules to specific to be reused

    Visual Programming Challenges

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    16/18

    End User Programming with agentsAgentSheets Study Report:

    AgentSheets is a production basedvisual programming language whereend users create with directmanipulation techniques

    Reports a study of teachers learningto build educational simulations ascurricula aids.

    Summary of results of formative

    evaluation to design agent basedproduction system for end usercreation of educational simulations.

    Proceedings of IEEE VisualLanguages 2002, Seals 2002.

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    17/18

  • 8/3/2019 2.ModelBased Reflex Agent Intro

    18/18

    Empirical Study Results

    Need robust drawing tools

    Objects should be important, not their spatiallocation

    Flexible object size

    Support for import of objects

    Allow incremental testing

    Increase the level of usability for noviceprogrammers

    Platform independent implementation