Re: IBL System Description and Architecture

Eric Jain wrote:

>> http://www.reengineeringllc.com/
>
>
> I intend to have a closer look at this, one day... Is there any 
> documentation available on the architecture of the system? If I 
> understand correctly, what the system basically does is generate SQL 
> statements from (more or less) natural language rules, right?
>
Hi Eric --

Yes, you are right in your description of the system, above.

For small amounts of data, the system can actually run without SQL -- 
see e.g. the demo called MedMine.

An important point is that the English vocabulary is open, and no 
dictionary or grammar maintenance is needed [1], so this is not 
yet-another-controlled-English system.

One can of course limit the vocabulary for the purpose of defining an 
ontology, but that is not a limitation of the system itself.

The system can produce hypertexted English explanations of results, or 
-- in the case of transactions -- of what it is about to do. This seems 
to be particularly useful when following complex chains of reasoning 
over RDF.

There's an overview architecture diagram on page 14 of the "Internet 
Business Logic Presentation" at www.reengineeringllc.com .

The "Semantic Web" and "Enterprise Modeling" presentations add some 
background.

HTH. Thanks in advance for comments. -- Adrian

[1] This statement often causes raised eyebrows among the natural language processing folks.  It's achieved via a trade off -- the rules are close to English, but a bit stilted, and the system never assumes that two different sentences mean the same thing unless it is explicitly told so.

---- 

Internet Business Logic  --  online at www.reengineeringllc.com

Reengineering LLC,  PO Box 1412,  Bristol,  CT 06011-1412,  USA

Phone 860 583 9677     Mobile 860 830 2085     Fax 860 314 1029

Received on Thursday, 19 August 2004 15:37:13 UTC