Monday, November 07, 2005

Monday 11:00-12:30 - lightning talks

Lightning talks start now...

Modelling overlaps between queries by counting the number of individuals that are members of both of two classes vs members of just one - Ven diagrammes. Looking arround, there are quite a few women - more than I was expecting. From Ven to bayes.

I totaly don't understand this seccond talk - q google search... http://cepa.newschool.edu/het/essays/uncert/vnmaxioms.htm

Next talk started with a picture of DNA - I feel better now. Garbled description of the central dogma, and a meaningless diagram. Cool. Language for recording experimental data and how well experimental data supports the ontology. I have a feeling that he has a hard seperation between the schema and the data.

Reasoning in the presence of inconsistencies - makes the point that the real world is almost certainly inconsistent, but with some logics inconsistencies are Very Bad Things. A question from the floor raised "contextual differences" as a source/highlighter of inconsistencies. Oh, and inconsistencies contain information, and ideally should not be discarded.

Another way to do probs in rdf that doesn't use quads. I think he's just rarified the model, which perhaps doesn't buy you anything.


0 Comments:

Post a Comment

<< Home