Professional Documents
Culture Documents
Knowledge-Based E-Business
Presented by: Thomas L. Adams, PhD
Mathematics & Computing Technology
Phantom Works
The Boeing Company
May 21, 2001
Briefing Outline
Trends in Cyberspace and Electronic Business
Technical Advances Impacting E-Business
Knowledge Representation in E-Business (Ontologies)
What is an Ontology?
Why Do We Need Ontologies?
Technical Issues in Using Ontologies
Ontologies for the Web (Semantic Web)
Ontology Standards
Ontology Projects at Boeing
Evolution of Cyberspace
HTML
Revolution
Shared Information
Space
Document Retrieval
Human Information
Processing
XML
Revolution
Culture of Open
Exchange
Dynamic Content Management
Business Service
Registries
Semantic
Web
Information and
Process Integration
Heterogeneous Information
Sources
Ad Hoc Query Answering
Integrated Knowledge
Management
and
Business Problem Solving
Distributors
Enterprise
Strategic
Partners
Service
Providers
Suppliers
Knowledge Representation
Highly Expressive Ontologies
Declarative Semantics
Multi-Agent Systems
Distributed Problem-Solving
Business Process Integration
Semantic Web
XML, XML Schema
RDF, RDF Schema
Topic Maps
Kinds of Ontologies
ad hoc
Hierarchies
(Yahoo!)
Terms
XML
Schema
structured
Glossaries
Thesauri
ordinary
Glossaries
Data
Dictionaries
(EDI)
DB
Schema
Principled,
informal
hierarchies
XML DTDs
formal
Taxonomies
Description
Logics
(OIL)
Data Models
(UML, STEP)
Frames
(OKBC)
General
Logic
Multi-dimensional continuum:
Increasing formality, amount of meaning specified, support for reasoning
Decreasing potential for amgibuity
Inference
M echanism s
Classes
M etaclasses
Slots/Attributes
Facets
Exceptions
Automatic classifications
Taxonomies
Inheritance
M onotonic, Non monotonic
Simple, M ultiple
Procedures
Execution of Procedures
Relations/Functions
Instances / Individuals / Facts
Axioms
Production Rules
Constraint Checking
Source: Evaluating Knowledge Representation and Reasoning Capabilities of Ontology Specification Languages,
Oscar Corho and Asuncion Gomez-Perez
Ontology Building
Maintenance
Validation and Verification
Merging
Evolution
Source: http://www.w3c.org
Ontology Languages
Enriched First-Order Predicate Calculus
CyC
KIF
Frame-Based Languages
Ontolingua
Frame Logic
Description Logics
Classic
OIL
What is Cyc?
Highly Expressive Knowledge Representation Language (CycL)
High Performance Inference Engine
Multiple Inference Strategies (about 30 currently)
Specification at Epistemological Level (EL)
Execution at the Heuristic Level (HL) based on syntactic
categories
Supports Default Reasoning (Arguments for and Against)
Knowledge is partitioned into microtheories
Very Large Repository of Common Sense Knowledge
Developers Toolkit
Ask, Assert, Create, Explanation
Concept Descripitons/Hierarchies
Java API (interaction with non-Cyc system elements
KE Text (batch processing of knowledge; testing)
UCEC
Defines attributes to describe products
Turbojet
engines
lift
thrust
starting
Propulsion
systems
engines
Engine
starters
Flame
propagation
Jet
engines
flameout
Flame
stability
Combustion
stability
Ramjet
engines
Hydrogen
fuels
combustion
afterburning
Burning
rate
Jet spray
spray
Uses of NCAS
Allow users to construct rather than select
concepts
Handle concepts not in the original Thesaurus
located in
Space
Information Access
tube placement
User
Agent
Metadata/
Ontology
Agent
User
Agent
Information
Broker
Agent
Information
Broker
Agent
Information
Service
Agent
Information
Service
Agent
Information
Service
Agent
Web pages
Databases
Experts
Metadata/
Ontology
Agent
Information
Service
Agent
Documents
...
FUTURE (?)
PRESENT
Clients
MAINFRAME
SOFTWARE
DATA
PROCEDURES
C
C
C
MAINFRAME
SOFTWARE
DATA
PROCEDURES
C
C
C
Server
Software
Object
Software
Object
Server
Software
Object
Software
Object
Server
Data
Object
Server
Data
Object
Server
Data
Object
Server
Data
Object
- Dumb Terminals
- Rigid Mainframe Applications
- Hard-Wired Connections
Procedural Software
Client-Server Architecture
- User Workstations (Clients)
- Application & Data Servers
- Programmed Connectivity
Object-Oriented Software
A
A
A
A
A
Centralized Architecture
A A
A
Server
Data
Service
A
A
Server
Data
Service
Server
Data
Service
Server
Data
Service
Agent Services
- Taskable User Agents
- Composable Network Services
- Dynamically Brokered Interaction
Agent-Based Software
- Autonomous Software Agents
- Dynamic Coordination Protocols
- Self-Organizing & Adapting
Autonomous
Agents cooperate to
achieve common goals.
Communication Protocols
Knowledge-Sharing
Coordination Strategies
Negotiation Protocols
Adaptive
Cooperative
Mobile:
Agents can either be static or mobile, depending on
bandwidth requirements, data vs. program size,
communication latency, and network stability
Market-Driven Economy
Team Coordination
Agents share knowledge about
goals, plans, tasks & subtasks,
commitments and performance.
Teams cooperative through
partially synchronized actions to
accomplish individual subtasks and
common goals (joint intentions).
Evolutionary Systems
Optional
Planner
Various
Capabilities
Conversation
Transport-Level
Support
Security Communication
Generic Agent
Generic
Agent
Instance
Agent A
Agent B
Agent-toAgent
Protocol
Generic
Agent
Instance
Autonomy
Business Units Have Different Goals, Preferences, and Resources
Goal-Directed Planning
Situation Monitoring/Rapid Response to Changing Conditions and Markets
Heterogeneity
Size (SME/Fortune 500)
Role (Buyer, Partner, Supplier)
National Differences (Culture, Legal Restrictions)
Dynamic Relationships
Rapid Formation/Dissolution of Business Relationships
Real-Time Negotiation of Contracts/Changes
Accommodate Cultural Differences in Styles of Interaction
Information
Brokering Agent
Resource
Descriptions
Local
Knowledge
Base
Negotiation
Agent
Planning
Agent
Scheduling
Agent
Integration
Agent
Translation
Agent
Adapted From: The Role of Agent Technology in Business to Business Electronic Commerce, Mike P. Papazoglou
5,7,10
1
B
3
1,12,14
1,6,9
8,11,13
D
2
Agent Standards
Foundation for Intelligent Physical Agents (FIPA)
Non-profit organization aimed at producing standards for the
interoperation of heterogeneous software agents
Main industry locus of action since 1996 for intelligent agents
knowledge-interchange standards work
Communication Primitives
Modes of Interaction (Query, Publish)
Interaction Protocols
Negotiation
FIPA Specifications
Guiding Principles
Openness (Dynamic Participation)
Agents can join or leave at run time without recompiling or
reconfiguring
FIPA naming conventions and registration process used to find the
location of another agent
Agent Management
Logical model for the creation, registration, location, communication,
migration and retirement of agents
Agent Communication
Communicative acts based on speech theory that are independent of the
content of the message
Interaction protocols describe entire conversations between agents for the
purpose of achieving some interaction of effect
Agent Applications
Service and ontology descriptions
Case scenarios
Integrating legacy software not communicating via FIPA ACL
Dynamic teaming
Brokering services
Variable semantics
Problem
User-Centric
Solving
Agents
Agents
Translation Agents
DARPA
http://www.aiai.ed.ac.uk/project/coax/
Policy-Mediated
Agent Interaction
What is a policy?
Formal specification to describe or govern interaction with other agents,
groups of agents, or grid resources
Provides a way to anticipate problems, facilitate and enforce desired
behavior, assist problem resolution, and justify sanctions for noncompliance
Examples
Security policies: e.g, all messages must be encrypted
Resource policies: e.g., ceilings on system resource consumption
Conversation policies: e.g., timing and message sequencing
Domain membership policies: e.g, agents cannot simultaneously be
members of domain X and domain Y
Benefits of explicit policy: reuse, operational efficiency, responsiveness to changed
conditions, possibility of off-line policy verification, tunable social control
Conclusions
Enrichment of Web-based Business Interactions with KnowledgeBased Technology Will Enable New eBusiness Capabilities
Shared Domain Conceptualizations (Ontologies) Provide the
Capability to Seamlessly Share Knowledge and to Identify Business
Partners and Opportunities
Rapid Advances in Expressivity and Inference Competency Can Be
Expected
Multi-Agent Systems Technology Provides the Capability for
Intelligent Interaction Among Diverse and Distributed Business
Entities
Interoperability of Heterogeneous Agent Frameworks and New Modes
of Interaction Are Under Intensive Development