Principals, Practice & Experience JS01 June 4-6, 2001 XML + Semantics = DARPA Agent Markup Language...

JS01 June 4-6, 2001

Principals, Practice & ExperiencePrincipals, Practice & Experience

XML + Semantics = DARPA Agent XML + Semantics = DARPA Agent Markup Language (DAML)Markup Language (DAML)

William Holmes, Dr. Paul Kogut

Management & Data Systems

Valley Forge, PA

June 4, 2001

JS01 June 4-6, 2001


RoadmapRoadmap The Semantic Web Agents & Ontologies Object Management Group (OMG) Initiatives The DARPA Agent Markup Language (DAML)

What is it? How does it fit in? / What is its role?

LM M&DS UML-based Ontology Toolset (UBOT) Ontology Design & Consistency Checking Automated Annotation via AeroTextTM

JS01 June 4-6, 2001


RING … RING ...Hello?

Hi Pete, it’s Lucy. I’m at the doctor’s office. Mom needs to see a specialist and then has to have a series of physical therapy sessions.

Biweekly or something. Can you split the chauffeuring with me?

Sure Lucy.

Semantic Web: The VisionSemantic Web: The Vision

Great! I’ll have my agent set up the appointments.

* Berners-Lee, Hendler, Lassila “The Semantic Web” Scientific American, May 2001

JS01 June 4-6, 2001


The VisionThe Vision

Semantic Web

Lucy’s agent retrieves informationabout Mom’s prescribed treatment

from the doctor’s agent.

Lucy’s agent looks up several lists of providers and checks for ones

in-plan for Mom’s insurance, within a 20-mile radius of her home, and

with a rating of excellent or very good.

Schedule a treatment plan for Mom using Pete and my schedules. Only use

providers that are in-plan for Mom’sinsurance, are within a 20-mile radius,

and have a rating of excellent or very good.

Lucy’s agent formulates a schedule of appointments for therapists with

appointments available that fit into Pete and Lucy’s schedule.

* Berners-Lee, Hendler, Lassila “The Semantic Web” Scientific American, May 2001

JS01 June 4-6, 2001


That’s Great but How?That’s Great but How? Need Agents

Definition (Merriam-Webster): one who is authorized to act for or in the place of another as a

business representative

Provide a means of processing the volumes of information found on the web.

Need Ontologies Definition:

Philosophy - A theory about the nature of existence. A.I. - A formal definition of relations among terms.

Provide a “semantic grounding” for the web.

JS01 June 4-6, 2001


What are Agents?What are Agents? In software, “Agent” is used in many different ways:

persistent process/daemon: mobile code autonomous robots “intelligent agent” - what makes it intelligent?

simple definitions that capture the essence of agents:

an Object that decides when to say go and when to say no - OMG

“programs that operate at a high enough semantic level that they can form new connections to other programs in order to get a job done” Burstein, McDermott

JS01 June 4-6, 2001


Why Agents?Why Agents? Agents are the next generation of middleware

– built on top of existing middleware (e.g., CORBA, EJB, Jini) – run-time integration via dynamic discovery and resource negotiation

– emphasis on broker and facilitator agents (e.g. yellow pages)

Agents are the next generation user interface– more complex applications require personal assistant agents – multi-modal interfaces e.g. speech, handwriting, gestures – user specifies goals and agent handles details according to user

preferencesInternet / Intranet agents

personal assistantagent

I need to go to Fort Worthon Monday for 3 days.

itinerary, tickets & maps

hotels car rental

airlinesmaps

JS01 June 4-6, 2001


Why Agents? (Cont.)Why Agents? (Cont.) Agents are the next level of component abstraction

agents are components with attitudes beliefs, desires, goals…*

agents interact like humans via speech acts request, inform, promise

agents share a context for efficient communication domain model ontologies are used at run-time ontology agent/services - query, retrieve and translate ontologies

*Labrou, Finin, Peng “Agent Communication Languages:The Current Landscape” IEEE Intelligent Systems March/April 1999

JS01 June 4-6, 2001


Examples of Agent Applications*Examples of Agent Applications* personal assistant - digital secretary

– travel arrangements– meeting schedule coordination– personalized information filtering– mobile computing

internet/intranet information retrieval/summarization electronic commerce enterprise workflow - e.g., sales, order processing, shipping military command and control synthetic characters (e.g., Extempo Systems, Virtual Personalities) robots - manufacturing, office, domestic design and engineering

*see Hendler “Is There An Intelligent Agent in Your Future?” http://helix.nature.com/webmatters/agents/agents.html

JS01 June 4-6, 2001


OntologiesOntologies Machine readable semantic specifications.

Include terms, relations, and inference rules What does “capital” mean?

Seat of government (Tallahassee, Harrisburg, Austin) An upper-case letter monies, securities, investments, etc… the top of a column or pillar.

XML is Not Enough!!! Allows definition of syntax, but not semantics (meaning) Can be considered the “Assembly Language” of the Web.

JS01 June 4-6, 2001


OMG InitiativesOMG Initiatives OMG Agent Platform Special Interest Group (SIG)

extend the OMG Object Management Architecture (OMA) to better support agent technology

identify and recommend new OMG specifications in the agent area recommend agent-related extensions to existing and emerging OMG

specifications promote standard agent modeling techniques see http://www.objs.com/agent/index.html

OMG Ontology Working Group Align the domain modeling activities of OMG with the Semantic Web

initiative of the World Wide Web Consortium and with related ontology development projects such as DARPA DAML and IEEE SUO (Standard Upper Ontology).

JS01 June 4-6, 2001


DDARPA ARPA AAgent gent MMarkup arkup LLanguageanguage Machine-Readable Ontologies & Annotation (markup) Aimed at “Resources”, Not just web-pages

Sensors Services Appliances

Lots of industry Buzz* Scientific American IEEE Distributed Systems New York Times ZDNet …

*See http://www.daml.org/inthenews.html

JS01 June 4-6, 2001


DAML: Basic IdeaDAML: Basic Idea

webpages

DAMLannotation

DAMLontologies

agentsagents

web crawlersweb crawlers

querie

s

querie

s

queriesqueries

linkslinks

annotateannotatemanually ormanually or

semi-automaticallysemi-automatically

DAMLannotation

RDBMS

data

schema

queriesqueries

queriesqueries

linkslinks

web pages, databases, legacy software, devices, sensors...web pages, databases, legacy software, devices, sensors...have annotations linking their terms to ontologieshave annotations linking their terms to ontologies

JS01 June 4-6, 2001


DAML Annotation: Extreme MetadataDAML Annotation: Extreme Metadata

documentdocumentparsing infoparsing info

Evolution of MetadataEvolution of Metadata

keywordskeywordsXMLXML

schemaschema

browserbrowser web crawlerweb crawler XMLXMLparsersparsers

Subject verb objectSubject verb objectsemantics forsemantics for

selected sentencesselected sentencesFull semanticsFull semanticsfor all content for all content

agentsagents(near-term)(near-term)

agentsagents(future)(future)

implicit semantic agreements on paper!implicit semantic agreements on paper!

explicit semantic agreements via machine-readable ontologiesexplicit semantic agreements via machine-readable ontologies

JS01 June 4-6, 2001


DAML ProgramDAML Program Main DAML website = www.daml.org Duration: August 2000 to Fall 2002 Approach:

MIT W3C semantic web activity http://www.w3c.org/2001/sw/ “The semantic Web and its languages” in IEEE Intelligent

Systems, November/December 2000, pages 67-73 available at http://www.ksl.Stanford.EDU/projects/DAML/

Extend XML/RDF represent ontologies annotate web pages and other information with links to ontologies

JS01 June 4-6, 2001


DAML Program DAML Program (Cont.)(Cont.)

17 research teams and 1 integration team industry, academia and World Wide Web Consortium expertise in AI knowledge representation, logic and web

technologies cooperation with European Union IST Program

www.daml.org/committee/

DAML language definition Ontology Definition Rules Definition

JS01 June 4-6, 2001


DAML Program DAML Program (Cont.)(Cont.)

DAML tools ontology development and verification web page annotation dynamic composition of agent services distributed query processing and inference ontology translation

DAML trial applications Government: Intelink, Center for Army Lessons Learned Commercial: e-commerce, information retrieval

JS01 June 4-6, 2001


The Origins of DAMLThe Origins of DAML Extensible Markup Language (XML)

provides syntactic interoperability depends on implicit semantic agreements

Resource Description Framework (RDF) designed to represent metadata for web resources in an XML syntax triples:

RDF Schema (RDFS) adds OO concepts: class and subclass

XMLXML

RDFRDF

RDFSRDFS

DAMLDAML

* For more information see www.w3.org

<shoeGen:GovermentOrganization rdf:ID="DARPA”/><shoeGen:OrganizationHomePage rdf:about="http://www.darpa.mil/"> <shoeProj:authorOrg rdf:resource="#DARPA" /></shoeGen:OrganizationHomePage>

JS01 June 4-6, 2001


Status of DAMLStatus of DAML DAML+Oil (ontology)

released January 2001 - latest revision March 2001 language specifications and documentation:

http://www.daml.org/2001/03/daml+oil-index.html

design rationale http://www.cs.man.ac.uk/~horrocks/Slides/index.html

DAML-L (logic) rule representation and reasoning development in progress

JS01 June 4-6, 2001


UML-Based Ontology Toolset (UBOT)UML-Based Ontology Toolset (UBOT) We are applying:

graphical modeling and formal verification techniques from software engineering

text extraction from natural language processing lexical semantic resources from cognitive science

to build a tool-set that supports creation, extension and consistency checking of DAML ontologies DAML annotation of information resources for agents

intended for users who have minimal training in knowledge representation and agent theory see http://ubot.lockheedmartin.com/

JS01 June 4-6, 2001


UBOT TeamUBOT Team Lockheed Martin Management & Data Systems

architecture, development and integration

Versatile Information Systems (Northeastern University) formal verification of UML

Lockheed Martin Advanced Technology Center field test of DAML and UBOT

Kestrel Institute automated formal methods

JS01 June 4-6, 2001


UBOT Architecture: Ontology EngineeringUBOT Architecture: Ontology Engineering

DAML Ontology Engineer

ExtendedDAMLontologies

XMImodels

UML GUI

UML Formalization

Specware

UBOT

Slangmodels

Consistencycheckingresults

Semanticinconsistencies

UMLDAMLTranslation

XMImodels

BaselineDAMLontologies

JS01 June 4-6, 2001


UBOT Architecture: AnnotationUBOT Architecture: Annotation

DAML Annotator

DAML annotated text or web pages

UML GUI

Text ExtractionText or

web pages

UMLDAML

Translation

Extractionto DAML

Translation

XMI

DAMLOntologies

automatically generated

uncorrected annotation

correctedannotation

UBOT

JS01 June 4-6, 2001


UBOT Architecture: COTS ComponentsUBOT Architecture: COTS Components UML GUI

Tau UML Suite (Telelogic)

Specware (Kestrel Institute) supports ontology consistency checking via formal methods SNARK theorem prover (SRI)

Text Extraction AeroText (LM M&DS)

extracts entities (e.g. people, organizations, etc.) from natural language recognizes relationships between entities

(e.g. [organization] hired [person] ) developed for the U.S. Intelligence Community 12 years experience with sophisticated linguistic processing many fielded applications

JS01 June 4-6, 2001


UML GUI: Tau UML SuiteUML GUI: Tau UML Suite

JS01 June 4-6, 2001


Text Extraction: AeroTextText Extraction: AeroText

Document Window

Extraction Display

JS01 June 4-6, 2001


Automatic Annotation: AeroDAMLAutomatic Annotation: AeroDAML

JS01 June 4-6, 2001


Questions?Questions?

Principals, Practice & Experience JS01 June 4-6, 2001 XML + Semantics = DARPA Agent Markup Language...

Documents

Transcript of Principals, Practice & Experience JS01 June 4-6, 2001 XML + Semantics = DARPA Agent Markup Language...