Intelligent Information
Systems on the Web
and in the Aether
Tim Finin
recommend
tell
register
University of Maryland
Baltimore County
Joint work with Scott Cost, Benjamin Grosof (MIT), Anupam Joshi,
Jim Mayfield (JHU), Charles Nicholas, Yun Peng, Yelena Yesha &
many students.
MAY 2002
tell
register
This work was partially supported by DARPA contract F30602-97-10215, NSF grants CCR007080 and IIS9875433 and grants from IBM,
Fujitsu and HP.
1
Overview




The Problem: building intelligent
information systems
The Semantic web as part of the
solution
Some work at UMBC
Comments and Conclusions
UMBC
an Honors University in Maryland
2
The problem




I’ve been engaged in research aimed at
developing intelligent information systems for
thirty years.
The problem is hard, progress is slow, but the
incremental results are worth it.
It’s a task for many generations to come.
Today’s environment is very different than that
in 1971.
UMBC
an Honors University in Maryland
3
They way we were…
UMBC
an Honors University in Maryland
AN IBM 360 circa 1971
4
They way we will be…
UMBC
an Honors University in Maryland
5
What’s new?





Internet. Virtually of the computers in the
world have been connected.
Scale. Every day many more computing and
communication devices are joining.
Power. Raw computing power continues to
climb.
Wireless. New technologies (GSM, 802.11,
Bluetooth, UWB?, IR, etc) are creating a
pervasive, ubiquitous computing environment
Web. The web is like Dennett’s "universal acid“,
a mythical chemical that eats through -- and
thus transforms -- everything in its path.
UMBC
an Honors University in Maryland
6
IDM Challenges
The environment makes new demands and offers
new challenges, enough to keep all of us busy, e.g.:
 Very open environments
 Large and diverse community of service and
content providers
 Lots of relative autonomy
 Dynamic ad hoc networks
 Systems with widely varying resources -bandwidth, connectivity, cpu, memory, disk,
power, software, knowledge, intelligence, etc.
UMBC
an Honors University in Maryland
7
Research topics
Concepts which can address these challenges
include:
Multiagent systems
 Information and knowledge sharing through common
representation languages, ontologies and protocols
 Automatic service description, discovery, composition
 Negotiation for services and information
 Trust based models for authorization, credibility and
security
 Social and norm governed behavior
 Delegation and degrees of autonomy
 Coordination and teamwork models

UMBC
an Honors University in Maryland
8
Semantic Web





I’ll argue that the semantic web provides a good
approach, language and tools to support the
development of intelligent information systems
in this environment.
This isn’t obvious, since the SW seems grounded
in the “traditional” wired web.
But, the principles which drive it are the right
ones for agents as well as pervasive computing.
And, by grounding agents in web technology,
they may make it out of the lab.
Next: overview of Semantic Web
UMBC
an Honors University in Maryland
9
Origins of the Semantic Web







Tim Berners-Lee’s original
1989 WWW proposal
described a Web of
relationships among named
objects that unified many
info. management tasks.
Guha designed MCF at Apple (~94)
XML+MCF=>RDF (~96)
RDF+OO=>RDFS (~99)
RDFS+KR=>DAML+OIL (00)
W3C’s SW activity (01)
W3C’s OWL (02?)
http://www.w3.org/History/1989/proposal.html
UMBC
an Honors University in Maryland
10
W3C’s Semantic Web Goals
 Focus
on machine consumption:
"The Semantic Web is an extension of the current web
in which information is given well-defined meaning,
better enabling computers and people to work in
cooperation." -- Berners-Lee, Hendler and Lassila, The
Semantic Web, Scientific American, 2001
current Web stores things whereas the
semantic Web does things.
 The
UMBC
an Honors University in Maryland
11
Semantic Web does what?





Concept-based search
 keyword-based search
Semantic navigation
 link-based navigation
Personalization
 one size fits all
Query answering
 document retrieval
Services
 CGI calls, but service-description languages,
negotiation, service composition, etc
UMBC
an Honors University in Maryland
12
Why is this hard?
This is what a web page looks like to a machine
And understanding natural language is not as hard as
understanding the images!
UMBC
an Honors University in Maryland
after Frank van Harmelen and Jim Hendler 13
OK, so HTML is not helpful
Could we tell the machine what the different parts of the
text represent?
name
education
CV
work
private
UMBC
an Honors University in Maryland
after Frank van Harmelen and Jim Hendler 14
XML to the rescue?
Some XML fans claim this could be done by adding
“meaningful tags” to parts of the text
< name >
< education>
< CV >
< work>
< private >
UMBC
an Honors University in Maryland
after Frank van Harmelen and Jim Hendler 15
XML  machine accessible meaning
But to your machine, the tags still look like this….
name >
< name
<education>
< education>
< CV
CV >
<work>
< work>
<private>
< private >
UMBC
an Honors University in Maryland
after Frank van Harmelen and Jim Hendler 16
Schemas take a step in the right direction
Schemas help….
< CV >
private
UMBC
an Honors University in Maryland
…by relating
common terms
between
documents
after Frank van Harmelen and Jim Hendler 17
But other people use other schemas
<name>
name >
<<educ>
education>
<> >
< work
<< CV
CV >>
<<>
private >
UMBC
an Honors University in Maryland
Someone else has one like this….
after Frank van Harmelen and Jim Hendler 18
The “semantics” isn’t there
< CV >
private
UMBC
an Honors University in Maryland
…which don’t fit in
after Frank van Harmelen and Jim Hendler 19
Ontologies can help …


An ontology defines the terms used to describe
and represent an area of knowledge.
Ontologies are used by people, databases, and
applications that need to share domain information (a
domain is just a specific subject area or area of
knowledge, like medicine, tool manufacturing, real
estate, automobile repair, financial management, etc.).
Ontologies include computer-usable definitions of basic
concepts in the domain and the relationships among
them ...
They encode knowledge in a domain and also
knowledge that spans domains.
In this way, they make that knowledge reusable.
UMBC
an Honors University in Maryland
Working Draft, Web Ontology Working Group.
20
Ontologies can help …
Catalog/ID
Thesauri
“narrower
term”
relation
Terms/
glossary
Simple
Taxonomies
UMBC
an Honors University in Maryland
Formal
is-a
Informal
is-a
Frames
(properties)
Formal
instance
Disjointness,
Inverse,
part of…
Value
Restriction
General
Logical
constraints
Expressive
Ontologies
After Deborah L. McGuinness (Stanford)
21
By providing “external” referents to merge on
SW languages add
mappings and structure.
nme
CV
CV
work
vate
educ
CV
ed
uc
UMBC
an Honors University in Maryland
after Frank van Harmelen and Jim Hendler 22
“The Semantic Web
will globalize KR,
just as the WWW
globalize hypertext”
TBL’s semantic web vision
-- Tim Berners-Lee
you are
here
UMBC
an Honors University in Maryland
23
Semantic web languages today



Today there are just two semantic web
languages
 DAML – Darpa Agent Markup Language
http://www.daml.org/
 RDF – Resource Description Framework
http://www.w3.org/RDF/
and one under development by the W3C
 OWL – Ontology Web Language
http://www.w3.org/2001/sw/
with more to come, IMHO
UMBC
an Honors University in Maryland
24
RDF is the first SW language
Graph
XML Encoding
<rdf:RDF ……..>
<….>
<….>
</rdf:RDF>
Good for
Machine
Processing
RDF
Data Model
Good For
Human
Viewing
Triples
stmt(docInst, rdf_type, Document)
stmt(personInst, rdf_type, Person)
stmt(inroomInst, rdf_type, InRoom)
stmt(personInst, holding, docInst)
stmt(inroomInst, person, personInst)
UMBC
an Honors University in Maryland
Good For
Reasoning
25
Simple RDF Example
dc:Title
http://umbc.edu/~finin/talks/idm02/
“Intelligent Information Systems
on the Web and in the Aether”
dc:Creator
bib:Aff
http://umbc.edu/
UMBC
an Honors University in Maryland
bib:name
“Tim Finin”
bib:email
[email protected]
26
XML encoding for RDF
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:dc="http://purl.org/dc/elements/1.1/"
xmlns:bib="http://daml.umbc.edu/ontologies/bib/">
<description about="http://umbc.edu/~finin/talks/idm02/">
<dc:title>Intelligent Information Systems on the Web and in the
Aether</dc:Title>
<dc:creator>
<description>
<bib:Name>Tim Finin</bib:Name>
<bib:Email>[email protected]</bib:Email>
<bib:Aff resource="http://umbc.edu/" />
</description>[email protected]
</dc:Creator>
</description>
</rdf:RDF>
UMBC
an Honors University in Maryland
27
N triple representation
 RDF
expressions can also be encoded as a set of
triples.
 <subject> <predicate> <object> .
<http://umbc.edu/~finin/talks/idm02/>
<http://purl.org/dc/elements/1.1/Title>
"Intelligent Information Systems on the Web and in the Aether" .
_:j10949 <http://daml.umbc.edu/ontologies/bib/Name> "Tim Finin" .
_:j10949 <http://daml.umbc.edu/ontologies/bib/Email> "[email protected]" .
_:j10949 <http://daml.umbc.edu/ontologies/bib/Aff> <http://umbc.edu/> .
_:j10949 <http://www.w3.org/1999/02/22-rdf-syntax-ns#type>
<Description> .
<http://umbc.edu/~finin/talks/idm02/>
<http://purl.org/dc/elements/1.1/Creator> _:j10949 .
<http://umbc.edu/~finin/talks/idm02/> <http://www.w3.org/1999/02/22rdf-syntax-ns#type> <Description> .
 Note
UMBC
an Honors University in Maryland
the generated ID for the anonymous node
28
Triple Notes

RDF triples have one of two forms:



<URI> <URI> <URI>
<URI> <URI> <quoted string>
Triples are also easily mapped into logic




<subject> <predicate> <object>
<predicate>(<subject>,<object>)
With type(<S>,<O>) becoming <O>(<S>)
Example:






subclass(man,person)
sex(man,male)
domain(sex,animal)
man(adam)
age(adam,100)
; Note: we’re not
; showing the actual
; URIs for clarity
Triples can be easily stored and managed in a DBMS
UMBC
an Honors University in Maryland
29
RDF Schema (RDFS)

RDF Schema adds
taxonomies for classes
and properties


and some metadata.


subClass and subProperty
domain and range
constraints on properties
Several widely used
KB tools can import
and export in RDFS
UMBC
an Honors University in Maryland
Stanford Protégé KB editor
30
RDFS supports simple inferences



An RDF ontology plus some RDF statements
may imply additional RDF statements.
This is not true of XML.
Example:
subproperty(mother,parent)
domain(parent,person)
range(parent,person)
Implies:
parent(eve,cain)
mother(eve,cain)
person(eve)
person(cain)
UMBC
an Honors University in Maryland
31
RDF is being used

RDF is being used in a number of W3C
specifications



Other web standards



CC/PP (Composite Capabilities/Preference Profiles,
http://www.w3.org/Mobile/CCPP/)
P3P (Platform for Privacy Preferences Project,
http://www.w3.org/P3P/)
RSS 1.0 (Rich Site Summary)
RDF calendar (~ iCalendar in RDF)
Other systems

Mozilla
UMBC
an Honors University in Maryland
32
RDF is not enough, but a good foundation
 RDF
lacks expressive adequacy for many tasks
Only range/domain constraints (on properties)
 No properties of properties (transitive, inverse etc.)
 No equivalence, disjointness, coverings, etc.
 No necessary and sufficient conditions
 No rules, axioms, logical constraints

 DAML+OIL
extends RDF
Layering makes partial knowledge available to
applications which only understand RDF
 NB: Building on RDF has some disadvantages

UMBC
an Honors University in Maryland
33
We’re going down a familiar road
KR trends

55-65: arbitrary data
structures
65-75: semantic networks
75-85: simple frame
systems
85-95: description logics

95-??: logic



Web trends

95-97: XML as arbitrary
structures
97-98: RDF
98-99: RDFS (schema) as
a frame-like system
00-01: DAML+OIL

02-??: DAML-L



Only much faster!
UMBC
an Honors University in Maryland
34
DAML+OIL as a Semantic Web Language
DAML
= Darpa Agent Markup Language
 DARPA program with 17 projects & an integrator
developing language spec, tools, applications for SW.
OIL = Ontology Inference Layer
 An EU effort aimed at developing a layered approach to
representing knowledge on the web.
Process
 Joint Committee: US DAML and EU Semantic Web
Technologies participants
DAML+OIL
 DAML+OIL specs released 01/01 & 03/01
 See http://www.daml.org/
 W3C SW activity started 08/01.
UMBC
an Honors University in Maryland
35
A Simple DAML Example
<rdfs:Class about="#Animal"/>
<rdfs:Class about="#Plant">
<daml:disjointFrom
resource="#Animal"/>
</rdfs:Class>

Note the mixture of rdf (plant and animal are classes) and
DAML (plant and animal are disjoint)
UMBC
an Honors University in Maryland
36
DAML+OIL  RDF
 DAML+OIL
ontology is a set of RDF statements
DAML+OIL defines semantics for certain statements
 Does NOT restrict what can be said

Ontology can include arbitrary RDF

But no semantics for non-DAML+OIL statements
 Adds

cardinality constraints, defined classes (=> classification),
equivalence, local restrictions, disjoint classes, etc.
 More

capabilities common to description logics:
support for ontologies
Ontology imports ontology
 But
not (yet) variables, quantification, and
general rules
UMBC
an Honors University in Maryland
37
DAML in One Slide
DAML is built on top of
XML and RDF
It allows the definition,
sharing, composition and
use of ontologies
DAML is ~= a frame
based knowledge
representation language
It can be used to add
metadata about anything
which has a URI.
URIs are a W3C standard
generalizing URLs
everything has URI
UMBC
an Honors University in Maryland
<rdf:RDF xmlns:rdf ="http://w3.org/22-rdf-syntax-ns#"
xmlns:rdfs="http://w3.org/rdf-schema#"
xmlns:daml="http://daml.org/daml+oil#“>
<daml:Ontology rdf:about="">
<daml:imports rdf:resource="http://daml.org/daml+oil"/>
</daml:Ontology>
<rdfs:Class rdf:ID="Person">
<rdfs:subClassOf rdf:resource="#Animal"/>
<rdfs:subClassOf>
<daml:Restriction>
<daml:onProperty rdf:resource="#hasParent"/>
<daml:toClass rdf:resource="#Person"/>
</daml:Restriction>
</rdfs:subClassOf>
<rdfs:subClassOf>
<daml:Restriction daml:cardinality="1">
<daml:onProperty rdf:resource="#hasFather"/>
</daml:Restriction> </rdfs:subClassOf> </rdfs:Class>
<Person rdf:about=“http://umbc.edu/~finin/">
<rdfs:comment>Finin is a person.</rdfs:comment>
</Person>
38
DAML-S




DAML-S is an ontology for describing properties
and capabilities of web services
Desiderata:
 Ease of expressiveness
 Enables automation of service use by agents
 Enables reasoning about service properties
and capabilities
Also appropriate for describing services in a
mobile/pervasive computing environment
See http://daml.org/services/
UMBC
an Honors University in Maryland
39
DAML-S components

Service profile (what it does)



Service model (how it works)




For service registration, discovery and matching.
High-level description of service and provider with a (human readable)
description of service, a specification of functionalities provided and other
functional attributes.
For service invocation, composition, interoperation, monitoring.
A service has inputs, outputs, preconditions and effects.
Composite processes are build using sequence, if-then-else, fork, etc.
Service grounding (how to access)

Specification of service access information (communication protocols,
transport mechanisms, etc.) which could be via SOAP, HTTP forms, Java
RMI, RPC, etc.
UMBC
an Honors University in Maryland
40
SW is work in progress


There are important language aspects which
need more work: rules, queries, etc.
Many tools need to be created



Applications need to be explored
The W3C is developing a new SW language


E.g., Protégé plug-in for DAML+OIL
OWL: Ontology Web Language
…
UMBC
an Honors University in Maryland
41
W3C Web Ontology
Working Group


The WOWG is working to create a
recommendation for the "Web Ontology
Language": OWL
51 Members from 30 W3C Organizations






Companies: Agfa, Daimler-Chrysler, EDS, Fujitsu, HewlettPackard, IBM, Intel, IVIS, Lucent, Network Inference, Nisus, Nokia,
Philips, Stilo, Sun, Unisys
Public Sector: DISA, Electricite de France, Intelink, INTAP, MITRE,
NIST
Research projects/Labs: DFKI, FZI, Ibrow group, Stanford, U.
Bristol, U. Maryland, U. Southhampton
Invited Experts: Medical, Digital Library, Defense, Technical
CoChairs: Jim Hendler, University of Maryland/MIND; Guus
Schreiber, Univ of Amsterdam/Ibrow
http://www.w3.org/2001/sw/WebOnt/
UMBC
an Honors University in Maryland
42
OWL Goals

The WOWG has identified the
following goals in developing OWL
 Shared ontologies
 Ontology evolution
 Ontology interoperability
 Inconsistency detection
 Balance of expressivity and scalability
 Ease of use
 XML syntax
 Internationalization
UMBC
an Honors University in Maryland
43
KR meets the Web



One way to think about the semantic
web is that we are creating a knowledge
representation language for the Web.
This is more than just selecting an appropriate
KR language and selecting an XML encoding.
The Web as an information system has many
significant properties.






Highly distributed
Many content providers
Dynamic
Evolving
Inconsistent
…
UMBC
an Honors University in Maryland
44
Semantic Web Principles

Everything is on the web


Partial information is assumed


Content and consensus is dynamic
Minimalist design


It’s not all true
Support information evolution


The web privileges scalability over integrity and there’s always
more and new stuff to find
Trust models are critical


People, places, times, things all have URIs
Make the simple things simple, and the complex things possible.
Standardize no more than is necessary.
Common data model

UMBC
an Honors University in Maryland
To support interoperability and knowledge sharing
Adapted from Eric Miller, W3C
45
Some UMBC applications
(1) Semantic web and agents (ITTalks)
(2) Information retrieval on the SW
(3) Service discovery and composition in
ad hoc mobile environments
(4) Distributed trust
UMBC
an Honors University in Maryland
46
(1) ITALKS
ITTALKS is a database driven web
site of IT related talks at UMBC and
other institutions. The database
contains information on
– Seminar events
http://ittalks.org/
– People (speakers, hosts, users, …)
– Places (rooms, institutions, …)
• Web pages with DAML markup are generated
• The DAML markup supports agent-based services
relating to these talks.
 Users get talk announcements based on the interests,
locations and schedules.
•
UMBC
an Honors University in Maryland
47
UMBC
an Honors University in Maryland
48
human
view
UMBC
an Honors University in Maryland
49
machine
view
UMBC
an Honors University in Maryland
50
ITTALKS Architecture
Web
Services
Web server +
Java servlets
People
HTTP
Email, HTML,
SMS, WAP
MapBlast, CiteSeer,
Google, …
HTTP, WebScraping
Apache
Tomcat
Agents
FIPA ACL, KQML,
DAML
SQL
DB
RDBMS
Databases
UMBC
an Honors University in Maryland
<daml>
<daml>
<daml>
</daml>
<daml>
</daml>
</daml>
</daml>
DAML
reasoning
engine
DAML files
51
ITTALKS Ontologies
 We’ve
defined and use the following ontologies,
all at http://daml.umbc.edu/ontologies/
calendar-ont.daml – calendar and schedule info
 classification.daml – ACM CCS topics
 person-ont.daml – people and their attributes
 place-ont.daml – talk locations
 profile-ont.daml – user modeling info
 talk-ont.daml – talks info
 topic-ont.daml – topics and interests

UMBC
an Honors University in Maryland
52
Two Advanced Capabilities
 I’ll
briefly describe two advanced
capabilities facilitated by DAML:
 Classifying talk topics and user interests
using DAML ontologies
 Using DAML as a communication
language among software agents
UMBC
an Honors University in Maryland
53
What are talks about?

Topic hierarchies provide
indexing terms

ACM CCS topic hierarchy

Open Directory
Encoded as DAML ontologies
 These allow users to specify interests as well as
browse the database of talks by topic
 Automatic classification of talks (based on title and
abstract) and users (based on his web pages, CV,

papers, etc.)

Discovery of mapping rules between CCS to OD
ontologies using IR techniques
UMBC
an Honors University in Maryland
54
Classifying Talks
e.g.: ACM
CCS
Topics Ontology
uses
ACM CCS Ontology
uses
CMU
Bow
statistical
text analysis
tools
Now is the
time for all
good men to
come to the
aid of the
country. Now
is the time for
ACM CCS
classifier
Training corpus
e.g.:5K ACM
abstracts
UMBC
an Honors University in Maryland
topics
55
Mapping between topic ontologies
T1
Training corpus T1
T2
UMBC
CMU
Bow
statistical
text analysis
tools
T1T2
mapper
Topic ontology T2
Training corpus T2
an Honors University in Maryland
t1:foo
Topic ontology T1
{(t2:bar, 0.8),
(t2:qux, 0.7),
…}
56
Interactive ontology mapper




Users create maps
between ontologies
with URIs to text
describing classes &
properties.
Automates mapping
process, taking into
account hierarchical
relationships and
user-specified
landmark mappings.
Text classification
used to compute
similarities between
pairs of classes or
properties.
A probabilistic
approach used to
combine hierarchical
information.
UMBC
an Honors University in Maryland
Used in XTalks to enable mappings between
Alternative topic ontologies in DAML+OIL
57
DAML and Agents
ask-all
subscribe
advertise

Much multi-agent systems work is grounded in Agent
Communication Languages (e.g., KQML, FIPA) and
associated software infrastructure such as the DARPA
Grid


The DAML program invites different paradigms which
will require some changes in ACLs and their associates
software systems.



The paradigm has been peer-to-peer message oriented
communication mediated by brokers and facilitators.
Agents “publish” beliefs, requests, and other “speech acts” on
web pages.
Agents “discover” what peers have published on the web.
The software agent research community is very
interested in the semantic web and DAML
UMBC
an Honors University in Maryland
58
10
1
ITTALKS app
mapquest
18
11
ITTALKS
agent
Travel
agent
17
user’s daml profile
Communication
protocol
FIPA ACL
9
2
API
12
User
agent
3
XSB
DAML+OIL
Reasoner
UMBC
an Honors University in Maryland
8
16
13 5
Calendar
agent
14
7
15
4
Broker
Agent
6
Agent
Name
Server
MS Outlook
Common agent infrastructure
MS Outlook
59
How does DAML Help?
interop
language
service
description
language
agent
communication
ontology
language
user
models
DAML+OIL provided a uniform language which met
Many needs in developing a complex application.
UMBC
an Honors University in Maryland
60
XTalks Personal Agent
External World
External
Plugins
COM
Bridge
Xtalks
Plugin
User
Interface
Mapquest
Plugin
User
Model
Buddy
List
Plugin
Personal
Plugin
Agent
Manager
Infrastructure
JADE platform
Rule
Engine
Interface
yajxb
XSB
Jess
XPA is a configurable “personal agent” which accepts
FIPA messages from XTalks and other instances of
XPAs as well as applications, e.g. MS Outlook.
UMBC
an Honors University in Maryland
61
Xtalks agents
Xtalks
Interface
Xtalks
System
FIPA Request
Response Protocol
Periodic
querying
Mapquest Agent
Personal
Agent (2)
Xtalks Agent
Personal
Agent (1)
UMBC
Xtalks
System
Scenarios
1,2
FIPA Request
Response Protocol
an Honors University in Maryland
1 – Xtalks Announcement
2 – User Agent Solicitation
3 – Buddy List
4 – Travel Planning
Scenarios
3,4
Personal
Agent (3)
62
Damlator translation engine




Extensible engine for DAML-encoded
Semantic Web pages translation and caching

Currently supported output formats:
 For humans: GIF and PNG
 For agents: DAML, NTriples, Prolog terms

Caching supports scalability and efficiency
Incorporated as an Apache-module

Faster, application/user independent and system-wide
availability

Accessed via http:[email protected]@/original/path/to/file.daml
HTTP
Similar to W3C RDF
Browser
Apache Web
DAMLATOR
Validation Service
Server
Module
DAML
Uses Jena RDF/XML Parser,
speaking
Agent
Apache Xerces, AT&T GraphViz
Available from
DAMLATOR
http://www.ittalks.org/download/
Local File System
Cache
UMBC
an Honors University in Maryland
63
How does DAML Help?
interop
language
service
description
language
agent
communication
ontology
language
user
models
DAML+OIL provided a uniform language which met
Many needs in developing a complex application.
UMBC
an Honors University in Maryland
64
(2) Integrating Retrieval and Inference

Problem: How do we do information retrieval over
documents and queries which combine free text and
semantic web markup?




IR systems and KB systems use different models
One Solution: (1) index both the text and markup and
then (2) use existing IR systems to find documents that
match queries
Issues: (1) How do we index markup? (2) When and
where do we do inferencing over the markup?
Applications: (1) Improved recall and precision for IR
systems, (2) Retrieving documents for question
answering.
UMBC
an Honors University in Maryland
65
Student Event Scenario


UMBC sends out descriptions of ~50 events a week to
students.
Each student has a “standing query” used to route event
messages.


Use LMCO’s AeroText system to automatically add
DAML+OIL markup to event descriptions.



A student only receives announcements of events matching his
interests and schedule.
Categorize text announcements into event types
Identify key elements and add DAML markup
Use JESS to reason over the markup, drawing ontology
supported inferences
UMBC
an Honors University in Maryland
66
Event Ontology
O rg a n ize r
E ve n t_ N a m e
...
EVENT
E ve n t_ D a te
S ta rt_ T im e
P la ce
E n d _ T im e
A simple ontology for
University events
 Includes classes,
subclasses,
properties, etc.
 Can include instance
data, e.g., UMBC,
NEC, Fairleigh
Dickenson, etc

DATE
T IM E
BA SEBA LL
...
T R IP
...
M O V IE
SH O W
SPO RT
TEAM
IN D IV ID U A L
...
BA SK ETBA LL
...
A T H L E T IC S
KEY:
In sta n ce O f
CLA SS
UMBC
an Honors University in Maryland
CH ESS
P ro p e rty
P ro p e rty
A sso cia tio n
S u b cla ss O f
67
IR Engine

We’re experimenting with two IR engines: JHU’s
Haircut and UMBC’s SIRE, using a similar
process for both:


Convert DAML markup to RDF triples
Infer additional triples which follow from model
(S,type,O) ^ (0,subclass,O2) => (S,type,O2)

Use domain specific rules to infer additional triples



“for a movie, retrieve genre property from IMDB”
Generate 7 indexing terms from each (S,P,O) triple
SPO, SP*, S*O, *PO, S**, *P*, **O
Index free text and resulting triple terms
UMBC
an Honors University in Maryland
68
HOWLIR FRAMEWORK
Movie
Sport
Talk
...
Event Categories
Event
information
in plain text
AeroText
+
Java
DAML/RDF
Markup
Generate
RDF Triples
Expand
Agents
Event
WEB
Description
RDF Triples
Trip
Inference with
DAMLJessKB
Expanded
RDF Triples
+ Free Text
Must
Query
User
Interface
Filter query on
event property
constraints
OK
Structured
Query
JHUHAIRCUT
IR Engine
Must
not
Events
Results User
Interface
UMBC
an Honors University in Maryland
Final Results
Inference on results
69
DOCUMENT
<DOC>
<DOCNO>'http://gentoo.cs.umbc.edu/howlir/announcements/charity#charity_001
</DOCNO>
<TEXT>'UMBC Blood Drive!!
Office of Student Life launches its annual Blood Drive for the Red Cross
on Mon, Nov 20 in the UC Ballroom from 10am - 4pm. </TEXT>
<TRIPLE>
triple(charity_001)(
'http://gentoo.cs.umbc.edu/howlir/announcements/charity#charity_001_place',
'http://daml.umbc.edu/ontologies/event_ont#Building',
'University Center').
triple(charity_001)(
'http://gentoo.cs.umbc.edu/howlir/announcements/charity#charity_001',
'http://daml.umbc.edu/ontologies/event_ont#Organizer',
'Office of Student Life').
triple(charity_001)(
'http://gentoo.cs.umbc.edu/howlir/announcements/charity#charity_001_date',
'http://daml.umbc.edu/ontologies/event_ont#Day_of_week',
'Monday'). … </TRIPLE>
</DOC>
UMBC
an Honors University in Maryland
70
QUERY
<Query>
<required>
triple(query_001)(
'http://daml.umbc.edu/ontologies/query#query_001’,
'http://daml.umbc.edu/ontologies/event_ont#Movie_Name'
'Ocean’s Eleven').
</required>
<allowed>
</allowed>
<disallowed>
triple(query_001)(
'http://daml.umbc.edu/ontologies/query#query_001’,
'http://daml.umbc.edu/ontologies/event_ont#Organizer’
‘SEB').
</disallowed>
</Query>
UMBC
an Honors University in Maryland
71
Results?


Doing experiments now to measure recall and
precision over a small collection of 1500 event
announcements and 12 queries.
Compare




Only free text
Free text + base triples but no inferencing
Free text + triples + inferred triples
We expect to see improved precision and recall
UMBC
an Honors University in Maryland
72
rty
rd
:s
rdfs
f
ro pe
so
las
rd f:P
(3) Enhancing Bluetooth’s
Service Discovery Protocol
S e rv ice
C
ub
f:P
ro
pe
rty
P ro v id e d
By
S e rv ice
C o st
A d H o c N e tw o rk
S e rv ice
rty
rope
rdf:P
O p e ra tin g
S ys te m
rd
fs
N e tw or k
T e c hn ol og y
:s
C o n ta ct
URI
ub
Cla
ss
O
P ro v id e r
N ame
f
P rin te rS e rv ice

P rin t
S pee d
Bluetooth’s SDP is very simple
 Services and attributes represented by UUIDs
which are 128 bit numbers!
 No registration, aggregation, multicasting, event
notification
Enhanced SDP uses DAML+OIL
 We assume at least one resource rich device in the
ad hoc network to serve as a matchmaker
 Services and attributes described in DAML using a
“standard” ontology
 All available information from service and attribute
descriptions used for matching
 Reasons to obtain closest match
UMBC
an Honors University in Maryland
P rio rity
V a lu e
ssO f

P rin tF ile
T yp e
u b C la
P rin tC o lo r
Q u a lity
y
ropert
rd fs:s
rdf:P
P rin te r
rdf:Pro
per ty
P ri nt
T e c hn ol og y
P rin t
O u tp u t
F o rm a t
P ri nt
R e so lu tio n
P rio rity
V a lu e
P rin te r
Mo del
73
(4) Delegation Based Model for Distributed Trust
 We
are developing a delegation based model
for distributed authorization and trust for use in
both wired and wireless scenarios.
 Trust depends on

policies + credentials + delegation actions + proofs
of permissions and obligations.
 Agents
make speech acts about and reason
over these properties and relations
 Grounded in an ontology represented in DAML
UMBC
an Honors University in Maryland
74
Other UMBC SW work






Context aware computing
Service composition in pervasive computing
environments
Intelligent opportunistic data caching in mobile
computing environments
Using DAML-S in FIPA’s directory facilitator
Ontology mapping
Better reasoning tools
UMBC
an Honors University in Maryland
75
Conclusions
 Some
thoughts…
 Solving the symbol grounding problem
 Rethinking agent communication
 How do we get there
UMBC
an Honors University in Maryland
76
The symbol grounding problem




An argument against human-like AI is
that it’s impossible unless machines
share our perception of the world.
A solution to this “symbol grounding
problem” is to give robots with human
MIT’s Cog
inspired senses.
But the world we experience is determined by our senses,
and human and machine bodies may lead to different
conceptions of the world (e.g. Nagel’s What Is It Like To
Be a Bat? )
Maybe the Semantic Web is a way out of this problem?
UMBC
an Honors University in Maryland
77
Solving the symbol grounding problem



The web may become a common world that both
humans and machines can understand.
Confession: the web is more familiar and real to
me than much of the real world.
Physical objects can be tagged with low cost
(e.g., $0.05) transponders or RFIDs encoding
their URIs
 See HP’s Cooltown project
http://cooltown.com/
UMBC
an Honors University in Maryland
78
Rethinking the agent communication paradigm

Much multi-agent systems work is grounded in
Agent Communication Languages (e.g., KQML,
FIPA) and associated software infrastructure.



This paradigm was articulated ~1990, about the same
time as the WWW was developed.
Our MAS approach has not yet left the laboratory yet
the Web has changed the world.
Maybe we should try something different?

The communication MAS paradigm has been peer-topeer message oriented communication mediated by
brokers and facilitators -- an approach inherited from
client-server systems.
UMBC
an Honors University in Maryland
79
Rethinking the agent communication paradigm
A possible new paradigm?
 Agents “publish” beliefs, requests, and other
“speech acts” on web pages.
 Brokers “search” for and “index” published
content
 Agents “discover” what peers have published on
the web and browse for more details
 Agents “speak for” content on web pages by


Answering queries about them
Accepting comments and assertions about them
UMBC
an Honors University in Maryland
80
How do we get there from here?




This semantic web emphasizes ontologies – their
development, use, mediation, evolution, etc.
It will take some time to really deliver on the
agent paradigm, either on the Internet or in a
pervasive computing environment.
The development of complex systems is basically
an evolutionary process.
Random search carried out by tens of thousands
of researchers, developers and graduate
students.
UMBC
an Honors University in Maryland
81
“The sheer height of the
peak doesn't matter, so
long as you don't try to
scale it in a single bound.
Locate the mildly sloping
path and, if you have
unlimited time, the ascent
is only as formidable as the
next step.” -- Richard
Dawkins, Climbing Mount
Improbable, Penguin
Books, 1996.
UMBC
an Honors University in Maryland
Climbing Mount Improbable
82
The Evolution of Useful Things




UMBC
an Honors University in Maryland
The Evolution of Useful
Things, Henry Petroski, 1994.
Prior to the 1890’s, papers
were held together with
straight pens.
The development of “spring
steel” allowed the invention
of the paper clip in 1899.
It took about 25 years (!) for
the evolution of the modern
“gem paperclip”, considered
to be optimal for general use.
83
So, we should …




Start with the simple and move toward the complex
 E.g., from vocabularies to FOL theories
Allow many ontologies to bloom
 Let natural evolutionary processes select the most
useful as common consensus ontologies.
Support diversity in ontologies
 Monocultures are unstable
 There should be no THE ONTOLOGY FOR X.
The evolution of powerful, machine readable ontologies
will happen over multiple human generations
 Incremental benefits will more than pay for effort
UMBC
an Honors University in Maryland
84
For more information

On RDF
http://www.w3.org/RDF/

On DAML
http://www.daml.org/

On W3C’s semantic web activity
http://www.w3.org/2001/sw/

On the semantic web
http://semanticweb.org/

On our work at UMBC
http://research.ebiquity.org/
UMBC
an Honors University in Maryland
85
UMBC
an Honors University in Maryland
86
Descargar

Document