[Attempto] Doubt regarding RACE

[Norbert E. Fuchs]

On 24 Apr 2012, at 8:22 , Ananya Poddar wrote:

> I have been trying to understand how to use the race web service.. Since I really want to integrate it with my proj..can you please help me?

After opening the RACE web-client (http://attempto.ifi.uzh.ch/race/) click on the button "Show Help" that will open the help window of RACE. At the top of this window you find the text "The web client is a front-end to the RACE web service. Read the RACE Web Service documentation to learn how to use RACE directly from your applications.". Click on the link "RACE Web Service documentation" (http://attempto.ifi.uzh.ch/site/docs/race_webservice.html) to get the description of the web-service that includes example calls and responses.

> As a part of our project, we have also generated a lexicon full of the words in our Database. I would like to ask, if there is anyway in which we can use our lexicon in RACE. While this facility is available in ACE, we have not been able to find it in RACE.

This is a very reasonable request that has come up before, and that needs to be taken care of once RACE goes into the public domain. Currently, however, I am busy extending RACE's reasoning power which will be needed in an EU project in which we are involved.

> And, also if there is any way in which we can directly feed our ACE parsable sentences as axioms in the RACE code, so that it is reasoned, and an answer is obtained for the query.

> Here, if there is any way whr our sentences can be fed as the axioms, as a back-end serving as a database of info. And then questions can be asked at the query area, thereby acting like an intelligent search engine.

RACE uses two different kinds of axioms to deduce theorems and to answer questions. First, there are domain-specific axioms that you feed as ACE sentences to RACE via its web-client or its web-service. Second, there are domain-independent auxiliary axioms that express general meaning postulates for the language ACE, e.g. the relation between plurals and singulars and the relations between natural numbers. These auxiliary axioms are expressed in Prolog, are RACE internal, and thus not accessible to the RACE users.

If I understand your request correctly, you are talking about domain-specific axioms. Why not define them once and for all in ACE, and feed them to RACE in addition to other ACE axioms?

Best regards.

Norbert E. Fuchs
Department of Informatics & Institute of Computational Linguistics
University of Zurich