Re: Apply Ontology Automatically

>>>>> "IBL" == Internet Business Logic <ibl@snet.net> writes:

  IBL> Phillip --

  IBL> You wrote (below) "ability ... to be able to apply the ontology
  IBL> automatically in some circumstances"

  IBL> This could be the major selling point.  Otherwise, the value of
  IBL> the ontology depends on how well programmers read, understand,
  IBL> and use it.  And, if they did that well, was it their
  IBL> value-add, not that of the ontology?

  IBL> Do you have examples in which an ontology has been applied
  IBL> automatically to do a significant real world task?


Whether you would consider the task to being significant or real
world, I don't know, but in this work we used an ontology to define a
the conclusions for a given set of results; the practical upshot of
this is that ontology terms get automatically applied to set of
proteins. This was work done mostly by Katy Wolstencroft at
Manchester. 


K. Wolstencroft, A. Brass, I. Horrocks, P. Lord, U. Sattler,
   R. Stevens, and D. Turi. A Little Semantic Web Goes a Long Way in
   Biology. In 4th International Semantic Web Conference, volume 3792,
   pages 786-800, Galway, Ireland, 2005.


To an extent this technique is limited because the relationship
between the data and the ontology is categorical -- the data can be
described as either one thing or another. In this work, by Keith
Flanagan at Newcastle, a similar thing is done, but there is a
probabilistic relationship between the data and the
conclusions. Obviously this is more powerful, but it adds complexity as
well. 

K. Flanagan, M. Pocock, R. Stevens, P. Lord, P. Lee, and
   A. Wipat. Logical and probabilistic reasoning for genomic
   rearrangement detection, 2005. Poster.

  IBL> (Questions intended constructively).

An taken as such. 


My apologies for only quoting work that I've been involved in; I'm
sure that there are many other examples. 

Phil

Received on Monday, 3 April 2006 16:20:26 UTC