[Attempto] DRS XML

[Marco Zanchi]

Ah yeah, I read about the RACE on a paper recently about using it with
ALEXA to do voice-activated logical reasoning ;)

It's good to have those sentence and token identifiers, I can probably use
them for error checking with an alternative tokenizer.

In general I find the APE pretty good, I feel really tempted to delve
deeper into the Prolog code to see if I can help too, but the
paradigm shift is quite radical for me and has been putting me off. It's
actually the first time I consciously see myself contemplating a
declarative language.

Thanks for your clarification.

On Tue, 17 Dec 2019 at 14:26, Norbert E. Fuchs <fuchs at ifi.uzh.ch> wrote:

>
>
> On 17 Dec 2019, at 5:51 , Marco Zanchi <mzanchi at gmail.com> wrote:
>
> I was reading the document Discourse Representation Structures for ACE
> 6.7, and currently working on creating a POJO framework for representing a
> DRS.
>
> The XML output from ACE has names that are a little different from this
> specification (like "struct" instead of "class" for example), but that's
> not a big issue. The reason for my message is that I don't find any
> reference on the document about the last two elements on a condition:
>
> <?xml version="1.0" encoding="UTF-8"?>
> <DRS domain="A B C">
>   <object ref="C" noun="father" struct="countable" unit="na" numrel="eq"
> num="1" sentid="1" tokid="2"/>
>   <object ref="B" noun="family" struct="countable" unit="na" numrel="eq"
> num="1" sentid="1" tokid="6"/>
>   <predicate ref="A" verb="belong" subj="C" sentid="1" tokid="3"/>
>   <modifier_pp ref="A" prep="to" obj="B" sentid="1" tokid="4"/>
> </DRS>
>
> What is "sentid" and "tokid"? My guess is that it's something to do with a
> tokenizer, but how so? Do we have a schema for this XML?
>
>
> Marco
>
> Your sentence
>
> The father belongs to the family.
>
> creates the DRS
>
> [A,B,C]
> object(C,father,countable,na,eq,1)-1/2
> object(B,family,countable,na,eq,1)-1/6
> predicate(A,belong,C)-1/3
> modifier_pp(A,to,B)-1/4
>
> where each DRS condition is labelled by sentid/tokid which indicates that
> the condition comes from the token tokid of the sentence sentid.
>
> This information is used, for example, in the Attempto reasoner RACE to
> identify the sentences of a set of axioms that are needed to prove the
> theorem. Here is an example.
>
>
> that shows that of the four axioms only axioms 2 and 3 are used for the
> proof.
>
> Best regards.
>
>    --- nef
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ifi.uzh.ch/pipermail/attempto/attachments/20191217/3794a361/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Screenshot 2019-12-17 at 8.22.45 .png
Type: image/png
Size: 88592 bytes
Desc: not available
URL: <http://lists.ifi.uzh.ch/pipermail/attempto/attachments/20191217/3794a361/attachment-0001.png>