Ontology Driven Systems for Search,

sounderslipInternet and Web Development

Oct 22, 2013 (3 years and 7 months ago)

61 views

Semantic (Web) Technology In Action:
Ontology Driven Systems for Search,
Integration and Analysis

Written by: Amith Sheth & Cartic
Ramakrishnan (IEEE bulletin, Dec 2003)

Presented by: Didit

Outline


Introduction


Semantic Web


Semantic Technology Application Nowadays


Ontology


Scalability Aspect in Ontologies


Semiformal, Less Expensive Ontology


Semagix Freedom: an Example of Semantic
technology


Conclusion


Introduction


Semantic web enable the web to
understand and satisfy the request of
people


Considered as the next phase of web
infrastructure development


How viable the web semantic is?

Outline


Introduction


Semantic Web


Semantic Technology Application Nowadays


Ontology


Scalability Aspect in Ontologies


Semiformal, Less Expensive Ontology


Semagix Freedom: an Example of Semantic
technology


Conclusion


Semantic Web


Provide standard based interoperability
of application


Semantic web vision


Representing knowledge


Acquiring knowledge


Utilizing knowledge


Scalability is the key challenge

Semantic Web


Closely supported by web service and
web process


Skepticism


related to lofty goal of the semantic web


related to scalability issue


Semantic Web


Scalability issue related to


Large ontology creation and maintenance


Large semantic annotation


Inference mechanism


Database technology play role in web
semantic


Involving in complex query processing



Outline


Introduction


Semantic Web


Semantic Technology Application Nowadays


Ontology


Scalability Aspect in Ontologies


Semiformal, Less Expensive Ontology


Semagix Freedom: an Example of Semantic
technology


Conclusion


Semantic Technology
Application Nowadays


Semantic search engine,

Taalee (Semagic)

[Townley
2000]


Provide domain specific search and contextual browsing


Extracting audio, video, and text content from 250 sources


Semantic integration, Equity Analyst Workbench
[sheth et al 2002]


Provide text content from site aggregated from 90
international source


Provide an integrated access to heterogeneous content.


Content are classified into small taxonomy and domain
specific metadata are automatically extracted

Semantic Technology
Application Nowadays


Analytics and knowledge discovery,
Passenger Threat Assessment [sheth et
al 2004]


Knowledge base is populated from many
public, licensed and proprietary knowledge
source


Periodic metadata extraction from
heterogeneous data is performed


Semantic Technology
Application Nowadays


Web semantic: review from existing
application


Application validate the importance of ontology in
the current approach


Ontology population is critical


Named entity and semantic ambiguity resolution is
important for data quality problem


Semi
-
formal ontologies (based on limited
expressive power) are most practical and useful


Large scale metadata extraction and semantic
annotation is possible

Semantic Technology
Application Nowadays


Web semantic: review from existing
application (continue)


Support for heterogeneous data is important
factor


Ability to query both ontology and metadata to
retrieve heterogeneous content is highly valuable


A majority of semantic web application
development rely on ontology creation, semantic
annotation, and querying

Outline


Introduction


Semantic Web


Semantic Technology Application Nowadays


Ontology


Scalability Aspect in Ontologies


Semiformal, Less Expensive Ontology


Semagix Freedom: an Example of Semantic
technology


Conclusion


Ontology


Ontologies come in
bewildering variety


Informal


Semiformal (do not use
formal semantic, the
ontology populated in
incomplete knowledge)


Formal


Semiformal ontology
demonstrated very high
practical value in term
of development effort


Outline


Introduction


Semantic Web


Semantic Technology Application Nowadays


Ontology


Scalability Aspect in Ontologies


Semiformal, Less Expensive Ontology


Semagix Freedom: an Example of Semantic
technology


Conclusion


Scalability Aspect in
Ontologies


Avalibility of large and useful ontologies


Must resemble as a database schema and capture
some aspect of real world semantic


Can be done through:


Social process (a suggestion and revision)


Automatically extract the ontology schema from
content


Automatic population of the knowledge base (with
respect to human design ontology schema)


Scalability Aspect in
Ontologies


Semantic metadata extraction of
massive content


Annotating heterogeneous content is very
challenging


The key to face this challenge is


Large scale availability of domain specific
ontologies


Scalable technique to annotate content

Scalability Aspect in
Ontologies


Inference mechanism that scale


Should can deal with massive number of
assertion


Deal with decidability and complexity issue
by limiting expressiveness of knowledge
representation

Outline


Introduction


Semantic Web


Semantic Technology Application Nowadays


Ontology


Scalability Aspect in Ontologies


Semiformal, Less Expensive Ontology


Semagix Freedom: an Example of Semantic
technology


Conclusion


Semiformal, Less Expensive
Ontology


Semiformal ontology more abundant and
useful than formal ontology


Can be built to a scale that is useful in real
world application


Can accommodate partial and possibly
incomplete information


The more semantic consistency required by
standard, the sharper the tradeoff between
complexity and scale


Outline


Introduction


Semantic Web


Semantic Technology Application Nowadays


Ontology


Scalability Aspect in Ontologies


Semiformal, Less Expensive Ontology


Semagix Freedom: an Example of Semantic
technology


Conclusion


Semagix Freedom: an Example of
Semantic technology


Provide modeling tool to design ontology
schema based on the application requirement


Exploits task and domain ontologies that are
populated with relevant facts


Perform


automatic classification of content


Ontology driven metadata extraction


Support complex query processing (semantic
search, semantic integration, and analytic
knowledge discovery)

Semagix Freedom: an Example of
Semantic technology

Semagix Freedom: an Example of
Semantic technology


Knowledge agent maintain ontology
automatically by traverse trusted source
periodically


Content agents have similar duty as
knowledge agents


Semagix Freedom: an Example of
Semantic technology


Semantic enhancement server enhanced incoming
content


identifying relevant document feature


Perform entity disambiguation


Tag metadata with relevant knowledge


Produce semantically annotated content


Support two form content processing


Automatic classification (utilize classifier committee based on
statistical, learning, and knowledge based classifier)


Metadata extraction (involve named entity identification and
semantic disambiguation)

Semagix Freedom: an Example of
Semantic technology


Metabase store both semantic and
syntactic metadata related to content


Semantic query server provide index of
metabase so that retrieval of metada
can be done quickly



Semagix Freedom: an Example of
Semantic technology


Performance capability


Typical size of an ontology schema: 10s of
classes, 10s of relationship, few hundred
properties type


Average size of ontology population: over a
million of named entities


Number of instance that can be extracted
in a day: up to a million per server per day

Semagix Freedom: an Example of
Semantic technology


Performance capability


Number of text document that can be processed
for metada extraction: hundreds of thousand to a
million per server per day


Performance for search engine type keyword
queries: over 10 million queries per hour


Query processing requirement in an analytical
application: approx 20 complex queries, taking
total of 1/3 second for computation

Outline


Introduction


Semantic Web


Semantic Technology Application Nowadays


Ontology


Scalability Aspect in Ontologies


Semiformal, Less Expensive Ontology


Semagix Freedom: an Example of Semantic
technology


Conclusion


Conclusion


Formal ontologies even though supported by
deductive inference mechanism may not be
the primary mean of addressing major
challenge in semantic web vision


Semantic web vision is not one of solving AI
problem


instead it can make critical contribution to the
semantic web (on issue in supporting
heterogeneous data, semantic ambiguity, complex
query processing, semiformal ontology, and ability
to scale large amount of information)