Transcript presentation slides contain more information

2004: A Semantic Web Odyssey
by Jens Bowman, Sudhir Poole, Boris 9000
OWL-DL
Motivation

The PL oracles said they will be…





not always available
unreliable
non-deterministic
imprecise (deliberately?!)
Therefore, as willing and eager PhD
students we decided to build…
The Tool

The Semantic (Web) Oracle!
Powered by
Main Research Hypothesis

Main problems in SW:

Difficult to access knowledge sources


Difficult to interpret results


Who wants to read RDF fragments?
Existing user interfaces suck!


unnatural query languages
We need a better one!
Our ideal:
Our 24 Hrs. (Research) Goals




Provide a query-answering system
based on natural text
Make Philipp jealous – we implemented
the system in 24 hrs. without having a
clue of what we’re doing
Make Nenad jealous – we developed in
24 hrs. new query refinement
techniques
Have fun!
Use Case
Who works at
AIFB?
Semantic Oracle
Architecture
Natural Language
Recognition Interface
Natural Language
Processor
Oxford’s
Dictionary of
Ancient Greek
Stop words insertion
?
Stemming
Part-of-speech-tagging
Multi-word term extraction
Discourse Representation
(in OWL, we are adding
F-Logic today evening)
Background
Knowledge
!
Text-to-Speech Processor
Answer Faking Module
Answer Generator
Multi-media Extensions
Demo!