Semantic Web without the Web

1
Semantic Web without the Web

• Sham Navathe
• Georgia Tech
• sham_at_cc.gatech.edu

2
Semantics of the web

• Semantics of the web
• - what is the meaning of a website?
• - what is the meaning of a webpage?
• - meaning of the page structure
• - meaning of the page contents
• - meaning of links

3
Research on semantic web

• Knowledge discovery
• Knowledge representation
• Knowledge understanding
• Knowledge integration
• Knowledge management

4
We did a lot of semantics research in 70s and
80s

• 1974 Abrial first paper on data semantics
• 1976 Chen - ER model semantics of
  entities/relationships/attributes (TODS)
• 1976 Navathe Schema Restructuring/mapping
  semantics across models Ph.D.thesis (TODS)
• 1977 Smith and Smith Abstraction of
  aggregation and generalization (TODS)
• 1979 Codd Extension of Relational Model to
  capture more meaning (TODS)

5
We did a lot of semantics work then

• 1981 McLeod- Semantic Data Model Ph.D. thesis
  semantics of aggregations/associations (TODS)
• 1985 Brachman Ph.D. KL-ONE Knowledge management
  System (Cognitive Science)
• 1987 Korth and Roth Nested Relational Model
  and Language to capture semantics of hierarchical
  nested data (SIGMOD)
• 1987 Hull and King Semantic Modeling a
  survey of research issues (Computing Surveys)
• 1988 Peckham and Maryanski Survey of Semantic
  Models (Computing Surveys)

6
We did a lot of semantics work then

• 1989 Beck/Gala/Anwar /Navathe- CANDIDE Data
  Model use of subsumption and description logic
  in query processing / use in conceptual
  clustering (ICDE - TKDE)
• 1989 Brachman et al.- CLASSIC system parallel
  work to above (SIGMOD)
• 1993 Sheth/Gala/Navathe Automatic Reasoning for
  Schema Integration (IJICIS)
• 1994 Navathe/Savasere et al. Using CANDIDE and
  subsumption based reasoning for integration (LNCS)

7
And a lot more on .

• Representation
• Knowledge extraction
• Mapping
• Integration
• Tools and systems express at ibm, dds at
  honeywell, infosleuth at mcc/bellcore, carnot at
  mcc

8
Going Forward .

• Semantic web community need not position itself
  as the discoverers of semantics
• Need to rely upon and acknowledge the work done
  before
• Semantic data models
• Knowledge representation
• Information Retrieval
• Reasoning
• Logic
• ..

9
Current State

• There seems to be general acceptance of work in
  A.I. And knowledge representation and reasoning
• Not enough acceptance of past work in database
  modeling, querying and metadata management
• Semantic Web has more sex appeal than
  database research which is obsessed by streaming
  data, mobile data management, olap, sensor queries

10
Current trends

• Knowledge is in, data is out
• Ontologies are in, schemas are out
• OWL and RDF are in SQL and Rel. Model are out

11
Hope

• The two communities will work together
• The semantic web folks will not consider them as
  owners of semantics
• The database folks will be curious enough about
  the exciting research in the semantic web
  community
• We start organizing more forums like this one to
  bring DB and semantics people together.