"Living" deliverable -
ENTER SOME EXPLAINATION/MOTIVATION HERE
Then start by collecting data sets and other requirements as subsections.
The attached document describes the structure of a NLS-GFT-OP corpus and its ingredients; the process of the creation of a NLS-GFT-OP corpus.
| Attachment | Size |
|---|---|
| creationOfNLS-GFT-OPICorpus2.pdf | 96.72 KB |
Please add your data set description as a sub page following the template format given here below.
Which data sets do you plan to use?
When adding your data set, cut and paste the sections here below.
Do you need any special software to create and edit the data set?
Describe some/all? queries that you need to execute against the set.
The modules and functionality of the MOLTO Knowledge Representation Infrastructure (KRI) is thoroughly described in D4.1 Knowledge Representation Infrastructure.
This KRI is the data modeling and manipulation backbone of the entire project, aiming to serve semi-automatic creation of abstract grammars from ontologies; deriving ontologies from grammars, and instance level knowledge from Natural Language (NL). In terms of retrieval, NL queries will be transformed to semantic queries and the resulting knowledge, expressed back in NL.
D4.1 Knowledge Representation Infrastructure can be found at: http://www.molto-project.eu/node/1120.
| Attachment | Size |
|---|---|
| D4.1.pdf | 518.37 KB |
scrolling="auto" frameborder="1">
[Your user agent does not support frames or is currently configured
not to display frames. However, you may visit
the related website.]
UGOT, with help and feedback from Ontotext, has started a GF grammar for queries. The grammar is currently for English only and translates queries like
"Bulgarian people working at Google"
to abstract syntax trees like
MQuery (QSet (SPlural (KProp (PRelation Employed (SInd (IName Google))) (KProp (PCountry Bulgaria) Person))))
The same syntax tree has hundreds of variants, such as
give me persons that are Bulgarian that are professionally active in Google give me Bulgarian people that work at Google give me people that are Bulgarian that work at Google Bulgarian people employed by Google
The grammar is very incomplete and unfinished, and very much work in progress! So contributions are welcome.
The code can be found in the GF darcs repository and viewed in http://code.haskell.org/gf/examples/query/ - see in particular README.
The requirements for the knowledge and representation infrastructure have been collected from the technological and use case partners. The technological partners with experience in semantic representations also brought in requirements based on previous experience and the state of the technology at the moment.