R: [SEM-GRD] towards the primer

["Mario Cannataro" <m.cannataro@libero.it>]

Hi Dave and Peter,

in our project Knowledge Grid (www.isi.cs.cnr.it/kgrid),
an environment for the design and deployment og grid-based distributed
data mining applications,
we are developing a two layer medatada system:

1) an ontology for the Data Mining domain, in DAML+OIL (RDF Schema) plus

     a DAML API-based tool for the browsing and querying of the
ontology,

that will be used to assist the designer of the data mining application
over of the Knowledge Grid in choosing the "best" data mining task,
methods, algorithms, and ultimately, the software and/or the suite to be
used; 
(we have developed the ontology and are developing the tool);

2) a set of metadata describing specific information about installed
data sources to be mined, data mining software,  extraction tools and so
on,

that is used when accessing the installed resources. Currently these
metadata are described in XML (we started the ontology project after the
development of layer 2).

Current work regards the development of a reasoning system over the
ontology,
in future we plan to introduce such an approach in the authoring system
of 
Grid-based Problem Solving Environments.
(regards PSE, we are interested in proteomic analysis, so I would be
interested to get in touch with people working on the field).

Thanks
Mario 

--
Prof. Mario Cannataro
Informatics and Biomedical Engineering, Faculty of Medicine,
University "Magna Gręcia" of Catanzaro,
Via T. Campanella, 115
88100 CATANZARO, ITALY
Tel: 0961-770880, Fax: 0961-777453
Email: cannataro@unicz.it, 
Web: www.isi.cs.cnr.it/isi/cannataro


>>-----Messaggio originale-----
>>Da: owner-sem-grd@gridforum.org 
>>[mailto:owner-sem-grd@gridforum.org] Per conto di David De Roure
>>Inviato: sabato 5 aprile 2003 14.05
>>A: Peter Vanderbilt
>>Cc: sem-grd@gridforum.org
>>Oggetto: Re: [SEM-GRD] towards the primer
>>
>>
>>Thanks - 
>>
>>Let me split that into two and throw it open...
>>
>>1. Example (RDF) metadata about data and/or compute resources, 
>>2. A deduction system and rules for that deduction system applicable 
>>to the metadata and an example conclusion.
>>
>>So, can anyone rise to this challenge?  Examples please!
>>
>>Re 1, there are many examples of catalogue-style metadata for 
>>data resources (cf dublin core) but I'm sure people can 
>>provide some richer examples.  For compute resources there 
>>are some specific examples I 
>>can think of including:
>>
>>- in the GRIA project we are expressing metadata relating to 
>>application specific feature vectors, workload estimation output 
>>vectors and capacity estimation parameter vectors
>>- the work of the scheduling ontology guys stems from a scheduling 
>>dictionary and maybe they can provide an example?  (Maybe 
>>something from GRIP?)
>>
>>The other metadata examples that spring to mind are about 
>>functionality 
>>of services rather than description of data/compute resources, but if 
>>we accept an OGSA world then the resources get wrapped by services so 
>>it blurs.
>>
>>If people can provide some concrete examples that would be great.
>>
>>Thanks
>>
>>-- Dave
>>
>>On Fri, 4 Apr 2003, Peter Vanderbilt wrote:
>>> David De Roure wrote:
>>> > [...]
>>> > 
>>> > Can I start by asking everyone on the list to tell me please what 
>>> > you would
>>> > like to see in the Semantic Grid Primer?  Imagine going 
>>to the web site - 
>>> > what would *you* expect to see there?
>>> 
>>> I would be very interested in information that would help 
>>me set up my 
>>> first semantic-grid-like system: for example, example (RDF) 
>>metadata 
>>> about data and/or compute resources, (a pointer to) a deduction 
>>> system, rules for that deduction system applicable to the 
>>metadata and 
>>> an example conclusion.
>>> 
>>> Or, if the above is not the sort of thing the semantic grid 
>>is about, 
>>> then I'd like to see some concrete examples of what it IS about. :-)
>>> 
>>> Pete
>>