IA by dandanhuanghuang

VIEWS: 5 PAGES: 72

									    Intelligent Agent
             David Yang
National Kaohsiung Normal University
     http://140.127.43.12/ia.htm
   Yangdav@nknucc.nknu.edu.tw
            Purpose of Presentation
• To overview the rapidly evolving area of
  software agents

• To present the real challenges and
  potential benefits of agents research




2002/3/16            Intelligent Agent       2
                Outline
• Background Introduction
• Historical Review
• Agent-Oriented Software
  Engineering(AOSE)
• Semantic Web & Agent
• Conclusion



2002/3/16         Intelligent Agent   3
             Video Show
• Knowledge Navigation (Apple, 1988)
• Computational AI is too limited and a
  controversial issue
• Weak AI (Searle, 1980): Computer robots
  can simulate human capabilities but will
  not replicate mentality
• Strong AI (Sloman,1992): AI systems will
  have mental states and processes

2002/3/16         Intelligent Agent          4
              Agents (Norvig, 1995)
                      percepts                   sensors


      environment                                          ?
                                                               agent
            actions

                 effectors

2002/3/16                    Intelligent Agent                         5
      A Unifying Theory of Arch.




•   perception,
•   – central states and processes,
•   – action mechanisms
•   (All with fuzzy boundaries)


2002/3/16                Intelligent Agent   6
                Another Common Arch.




•    Evolutionary age (reactive oldest from brain research
•    – Level of abstraction of processing
•    – The types of control functions, and mechanisms used
•    – The forms of representation used

    2002/3/16              Intelligent Agent            7
 CogAff Schema(Sloman,2001)




• CogAff :defines only possible components and links –
  not all need be in all organisms, or all robots
• An ‘ecosystem’ of mind: A grid of co-evolving sub-
  organisms each contributing to the niches of the others.


 2002/3/16               Intelligent Agent             8
       Why do we need the intelligent
            agent technology
• Information explosion and overload due to
  the popularity of internet technology
      – information supply (provider)
      – Information demand (consumers)
• Personal assistance of information access
  and utilization



2002/3/16              Intelligent Agent      9
       Why do we need the intelligent
         agent technology(Cont.)
• Enhance efficiency of communication and
  interaction among people
• Construction of virtual agent-based
  communities and societies




2002/3/16          Intelligent Agent        10
            What is Software Agent?
• Particular type of agent
• Inhabiting computers and network
• Assisting users with computer-based tasks

                                                     User
                       Interacts with

        Applications                                    collaborate

                        Interacts with               Agent




2002/3/16                        Intelligent Agent                    11
             Software Agents: History
• Agents research dates from the early days
  of AI work

• Carl Hewitt’s actor model (1997)
            “a self-contained, interactive and
            concurrently-executing object, with some
            encapsulated internal state and which could
            respond to messages from other similar
            objects”
2002/3/16                    Intelligent Agent            12
            Software Agents: History


              Context of agents research


                       Distributed AI


    Distributed          Agents and          Parallel AI
  Problem Solving   Multi-agent Systems



2002/3/16                Intelligent Agent                 13
 Motivation for Agents Research
• Benefits from DAI

       Modularity which reduces complexity

       Speed due to parallelism

       Reliability due to redundancy

       Flexibility new tasks are composed more
        easily from the more modular organisation

2002/3/16                Intelligent Agent          14
 Motivation for Agents Research
• Benefits from AI

       Operation at the knowledge level

       Easier maintenance

       Reusability

       Platform independence


2002/3/16               Intelligent Agent   15
               Software Agents:
            Strand One: 1977- date
• Concentrated on deliberative-type
    agents with symbolic internal models

• Addressed macro issues emphasizing the
  society of agents over individual agents

• Had the goal to specify, analyze, design
  and integrate systems comprising of
  multiple collaborating agents
2002/3/16           Intelligent Agent        16
            Strand 1 Agents Work:
                macro issues
• The interaction and communication
  between agents (Speech Act Theory)

• The decomposition and distribution of
  tasks among agents

• Coordination and cooperation

• Conflict resolution via negotiation

2002/3/16           Intelligent Agent     17
             Software Agents:
            Strand Two (1990 )

• research and development of agent
  theories, architectures and languages

• a significant broadening of the typology of
  agents being investigated



2002/3/16           Intelligent Agent           18
            Agent Applications
• workflow management          • digital libraries
• network management           • personal digital
• air-traffic control            assistants
• business process             • diary management
  engineering                  • e-mail filtering
• smart databases              • information
• command and control            management
• education                    • data mining
                               • electronic commerce


2002/3/16           Intelligent Agent                  19
            Agent Applications


    “in 10 years time most new IT
    development will be affected, and many
    consumer products will contain embedded
    agent-based systems” (Guilfoyle 1995)



2002/3/16           Intelligent Agent     20
            Don’t constantly need       Are able to work
            instructions                    unaided
                                Proactive




       Share
  information with                                       Improve their
     each other.                                          actions with
                 Cooperate                         Learn   experience
  Able to agree on
      subtasks

                     Software Agents
2002/3/16                      Intelligent Agent                         21
                Typology of Agents

                                              Collaborative
                                               Collaborative
                                             Learning Agents
                                             Learning Agents
             Cooperate
             Cooperate              Learn
                                    Learn

Smart
 Smart
Agents
Agents
                   Autonomous
                   Autonomous
                                              Interface
                                               Interface
Collaborative
Collaborative                                  Agents
                                                Agents
   Agents
   Agents
 2002/3/16               Intelligent Agent                 22
   Classification of intelligent agents

• Classify by mobility
      – Stationary
      – Mobile
• Classify by architecture
      – Deliberative
      – Reactive
      – social



2002/3/16              Intelligent Agent   23
            Classification of intelligent
                   agents(Cont)
• Classify by attributes
      – Collaborative agents
      – Interface agents
      – Learning agents
• Classify by roles
      – Match marker
      – Information gathering
      – Tutorial agents, etc

2002/3/16                Intelligent Agent   24
            What is an Agent?
• A component of software and/or hardware
  which is capable of acting exactingly in
  order to accomplish tasks on behalf of its
  user

• An umbrella term which covers a range of
  other more specific agent types


2002/3/16          Intelligent Agent         25
  Characteristics of software agents

• Proactive: goal-oriental
• Autonomy: bounded rationality
• Social: communication, cooperation,
  team/coalition formation
• Adaptive: learning from environment and
  other agents
• Mobility: interoperability, navigation,
  persistent
2002/3/16         Intelligent Agent         26
            What’s new in software agent
                    technology
• A paradigm shift of information utilization
  from direct manipulation to indirect access
  and delegation
• A kind of middleware between information
  demand (client) and information supply
  (server)
• A software that has autonomous,
  personalized, adaptive, mobile,
  communicative abilities

2002/3/16              Intelligent Agent    27
 What intelligent agents can do?
•   Internet data gathering and retrieval
•   Electronic news and mail filtering
•   Calendar management and meeting scheduling
•   Work-flow assistants
•   Making travel arrangement
•   Event monitoring
      – Alerting users to investment opportunities
      – Alerting doctors on emergency events of patients


2002/3/16                  Intelligent Agent               28
            Typical examples for human
                      societies
• Electronic commerce
      – Buyer, sellers, brokers, banks…..
• Virtual enterprise
      – Managers, manufacturers, engineers,
        accountants
• Digital library
      – Librarians, users



2002/3/16                   Intelligent Agent   29
            Typical examples for human
                  societies(Cont.)
• Virtual school
      – Teachers, students, classmates, staff
• Virtual hospital
      – Physicians, staff, patients
• Electronic government
      – Policy makers, officers, policemen, citizens




2002/3/16                 Intelligent Agent            30
  Impacts of the software agents on
        information services
• Information services become interactive
  and personalized
• The service provider can record and
  induce from the user’s behaviors
• The service provider can easily customize
  its services to different customers with low
  cost


2002/3/16           Intelligent Agent        31
  Impacts of the software agents on
     information services(Cont.)
• Enhance the quality
• Reduce the cost of information services
• Provider integrated services




2002/3/16          Intelligent Agent        32
            Environment of agents
•   Accessible versus inaccessible
•   Deterministic versus nondeterministic
•   Episodic versus nonepisodic
•   Static versus dynamic
•   Discrete versus continuous




2002/3/16            Intelligent Agent      33
    How AI techniques used in the IA
               domains?
•   Knowledge representation and inference
•   Planning
•   Machine learning/data mining & discovery
•   Common sense reasoning
•   Uncertainty reasoning




2002/3/16           Intelligent Agent          34
   How AI techniques used in the IA
           domains?(Cont.)
• Distributed (multiple agents) problem
  solving
• Distributed Constraint Satisfaction
  Problem (DCSP)
• Intelligent man/machine interface
• User modeling--personalized



2002/3/16          Intelligent Agent      35
         Comparison between expert
        systems and intelligent agents
• Expert systems address less
  communication and coordination issues
• Expert systems tend to knowledge-driven
  and address less inferring joint intention
  and team behaviors
• Address nothing about mobility
• Autonomous agents versus consultant
  through dialogue

2002/3/16           Intelligent Agent          36
       Types of agent Architecture
•   Logic-base agents
•   Reactive agents
•   Belief-Desire-Intention(BDI) agents
•   Layered architectures




2002/3/16            Intelligent Agent    37
     Process of Simple Reflex Agent
 function SIMPLE-REFLEX-AGENT( percept) returns
    action
  static: rules
  state ←INTERPRET-INPUT(percept)
  rule ←RULE-MATCH(state, rules)
  action ←RULE-ACTION[rule]

 return action


2002/3/16            Intelligent Agent            38
            BDI Architecture
                                     1. function action( p : P):A
                                     2. begin
                                     3.   B :=br f( B,p)
                                     4. D :=options(D,I);
                                     5. I :=f ilter( B,D,I);I
                                     6. return execute( (I);
                                     7. end function action



2002/3/16        Intelligent Agent                              39
            Layered Architectures




2002/3/16           Intelligent Agent   40
   Agent-Oriented
Software Engineering
            Research Topics on AOSE
• UML and agent systems
• Agent-oriented requirements analysis and
  specification
• Software development environments and CASE
  tools for AOSE
• Standard APIs for agent programming such
  JDK1.5
• Formal methods for agent-oriented systems,
  including specification and verification logics

2002/3/16            Intelligent Agent          42
            Research Topics on AOSE
                     (Cont.)
• Engineering large-scale agent systems
• Experiences with field-tested agent
  systems
• Best practice in agent-oriented
  development
• Market and other economic models in
  agent systems engineering
• Practical coordination and cooperation
  frameworks for agent systems
2002/3/16            Intelligent Agent     43
  Knowledge Representation for the
     “Semantic Web” with Agent


                                         DAML-L (Logic)

                                 DAML + OIL (Ontology)

                            RDFS (RDF Schema)

             RDF (Resource Description Framework)

            XML (Extensional Markup Language)




2002/3/16            Intelligent Agent                    44
               The Evolving Web
 • Locating Resources
    – free text & keyword search  semantic search
 • Web Users
    – primarily humans  both humans and machines
 • Web Tasks & Services
    – a place to find things  a place to do things

 Semantics is the Core Requirement
       – web content with no semantics  with semantics




2002/3/16                  Intelligent Agent              45
   Agents and the Semantic Web
 • Semantic Web: killer ‘app’ for agents?
 • Agents need to communicate and understand
   meaning.
    – Advertise and require capabilities
    – Locate meaningful information resources on
      web & combine them in meaningful ways to
      perform tasks
    – How to interpret communication acts?
 • But what do we mean by the Semantic Web?

2002/3/16             Intelligent Agent            46
             What do we mean by
                ‘Semantics’?
• Semantics of What?                                  Implicit
   – language? term? expression?
   – communication protocol?
   – domain ontology & markup!
                                                      Informal
• Plicity: Are the semantics implicit or explicit?
• Formality: How are semantics expressed?
• Semantics Processing: Who are they for?
                                                      Formal
   – human only – fully manual
                                                     Comments
     – human and computer – partially automated
     – computer only – fully automated
                                                     Automated
 2002/3/16                 Intelligent Agent               47
                    Examples
• Implicit: based on human consensus, shared
  understanding
   – Typical XML tags
       • <price>          200      </price>
       • <address>         …       </address>
       • <delivery-date> … </delivery-date>
   – Used by screen-scrapers, wrappers
   – Rife with ambiguity.
• Informal: only humans can use (until NLP solved)
   – Text specification document for HTML e.g. <h2>
   – UML semantics document
   – Java language definition, for compiler writers
   – Still ambiguous
 2002/3/16              Intelligent Agent             48
                      Examples
 • ‘Formal Comments’
     – Semantics of FIPA ACL ‘inform’ in modal logic
     – Formal definitions in any requirements spec (e.g. Z)
     – Many axioms in Ontolingua ontologies
     – Much less ambiguous
     – Still error-prone, human in the loop.
 • Automated
     – RDF(S), DAML+OIL term definitions
       e.g. mammal, date
       – How does the machine process the semantics?



2002/3/16                 Intelligent Agent                   49
  Machine Processible Semantics
• How can an agent learn the meaning of a term?
• Procedural Semantics
   – How does an agent system know what to do when it sees
     the term ‘inform’
   – The (possibly informal) semantics of ‘inform’ is
     embedded in a procedure by a human.
   – The system places a call to the procedure when it
     encounters ‘inform’.
   – The ‘meaning’ of ‘inform’ is what happens when this
     procedure is called.
• Machine processible semantics? – perhaps.



 2002/3/16               Intelligent Agent             51
  Machine Processible Semantics
• Learning the meaning of a term from a formal
  declarative specification of the semantics…
• General case: no assumptions, nothing shared
   – all symbols might as well be in ‘Greek’ script
   – no knowledge of language syntax, or semantics
   – Cryptography, impossible to automate
   – So, we have to cheat…
• We must make some assumptions…


 2002/3/16            Intelligent Agent           52
             Assumptions: Language
• Shared language syntax and semantics,
   – e.g. KIF, RDF(S), DAML+OIL


• But: may have incompatible assumptions in
       conceptualization.
   – Time point, vs. time interval
   – Agent can never incorporate meaning of new
     term in its axioms.


 2002/3/16            Intelligent Agent           53
              More Assumptions:
                Compatibility
• Logical compatibility as well as language.
• But: Different people build different ontologies
        for the same domain.
   – Two terms, same meaning, or vica versa;
   – Same concept modeled at different level of detail;
   – Different language primitives used for same concept;
       • e.g. red an attribute, or RedThings a class.
• Computationally intractable to determine if two terms actually
  mean the same thing.
   – I.e. have same set of models


 2002/3/16                 Intelligent Agent                 54
         More Assumptions: Sharing
• Term explicitly mapped to a shared concept
   – Encounter new term, leprechaun, a subclass of mammal.
   – ‘mammal’ defined in shared animal ontology in OIL.
• Machine can learn something about meaning.
   – I.e. there are now more things that it cannot be.
   – Still plenty of scope for ambiguity;
   – Definition of mammal in OIL can never be complete.
• Can do some inference
   – e.g. for search application looking for content about
     mammals.


 2002/3/16               Intelligent Agent             55
             Processing Semantics
• Relies on a formal semantics of OIL to infer semantics
  of terms and expressions in OIL.
• OIL semantics is for humans
   – it helps build inference engines;
   – not machine processible.
• Humans may still embed some meaning in code
   – May be dangerous to do so – or –
   – May be necessary to do so…
• The shared concept referred to may not be formally
  defined (e.g. Dublin Core terms)



 2002/3/16               Intelligent Agent             56
            Definition of Ontology
• An ontology is a formal, explicit
  specification of a shared conceptualization.
      – Formal
      – Explicit specification
      – Shared
      – conceptualization




2002/3/16                 Intelligent Agent   57
      What is a conceptualization?
                         a

                         b               d

                         c               e




                    Scene 1: blocks on a table

            Conceptualization of scene 1:
            <{a, b, c, d, e }, {on, above, clear, table }>


2002/3/16                     Intelligent Agent              58
What is a conceptualization?
                                         c

                        a               d

                        b               e




            Scene 2: a different arrangement of blocks


                The same conceptualization?


2002/3/16                   Intelligent Agent            59
What is a conceptualization?
• Conceptualization: the formal structure of
  reality as perceived and organized by an
  agent, independently of:
      – the vocabulary used (i.e., the language used)
      – the actual occurrence of a specific situation
• Different situations involving the same objects,
  described by different vocabularies, may share
  the same conceptualization.
            LE   apple
                                         same conceptualization
            LI   mela
2002/3/16                   Intelligent Agent                     60
  An ontology model for FIPA agent
            applications
• Frames
      – Frames are used to represent concepts.
• Relations
      – KIF relations, are used to represent concepts
        we will simply described by a value and to
        define concepts as combination of other
        concepts.



2002/3/16                Intelligent Agent              61
    What are we/agents going to use
            ontologies for?
• To share common understanding of
  conceptual models
• To enable knowledge re-use
• To make assumptions about
  knowledge/domain models explicit
• An ontology describes the concepts and
  relationships that can exist for an agent or
  community of agents

2002/3/16           Intelligent Agent            62
              Human’s Talking
• I am a graduate student at GIAE in NKNU.
• Keyword:
      – student
      – graduate
      – GIAE
                                       Point ?
      – NKNU




2002/3/16          Intelligent Agent             63
    Domain and scope of the ontology
Domain : Student
Scope : Classmate of senior high school
Class student {
   String School;
   String Institute;
   String Grade;

    public String getSchool() { return School; }
    public void setSchool(String s) { School = s; }
    public String getInstitute() { return Institute; }
    public void setInstitute (String d) {Institute = d; }
    public String getGrade() { return Grade; }
    public void setGrade (String g) { Grade = g; }
}

2002/3/16                          Intelligent Agent        64
        Reusing existing ontologies
•   Ontolingua ontology library
•   DAML ontology library
•   UNSPSC(United Nation)
•   RosettaNet(Communication)
•   DMOZ(Medicine)




2002/3/16           Intelligent Agent   65
            Ontology Agent




2002/3/16        Intelligent Agent   66
        Enumerate important terms
• Educational resources
      – IEEE’s Learning Technology Systems
        Architecture (LTSA)
      – Metadata
            • Dublin Core (DC) Metadata
            • Learning Objects Metadata (LOM)
      – SCORM Metadata
            • IEEE LTSC Learning Objects Metadata
            • IMS Learning Resource Metadata XML Binding
              Specification

2002/3/16                    Intelligent Agent             67
        Important Lesson Plan-related
                    terms
• Course’s catalog
• Catalog’s entry
• A entry’s title, format, location,
  typicallearningtime…
• Entry’s relation
      – Kind
      – resource


2002/3/16            Intelligent Agent   68
     Define the classes and the class
                 hierarchy
• approaches:
      – Top-down
      – Bottom-up
      – Combination
• Note : Only define the classes




2002/3/16             Intelligent Agent   69
   Define the properties of classes--
                 slots
• Intrinsic
• Extrinsic
• Relationships to other individuals




2002/3/16          Intelligent Agent    70
            Define facets of the slots
• Slot cardinality
• Slot-value type
      – String
      – Number
      – Boolean
      – Enumerated




2002/3/16             Intelligent Agent   71
            Conversion




2002/3/16      Intelligent Agent   72
             Conclusion
• The development process of Intelligent
  agents is evolutionary
• Agents combine and force with semantic
  web for information retrieval
• Agent Communication Language(ACL)
  based upon KQML (Knowledge Query
  Markup Language)
• Ontology focuses on conceptual
  development beyond logics

2002/3/16         Intelligent Agent        73

								
To top