Section: Software


Participants : Bruno Guillaume [correspondant] , Guy Perrier.

Graph rewriting, Interface syntaxe-sémantique

Grew is a Graph Rewriting tools dedicated to applications in NLP. It is freely-available (from the page http://grew.loria.fr ) and it is developed using the InriaGforge platform (http://gforge.inria.fr/projects/semagramme/ )

We list below some of the major specificities of the GREW software.

  • Graph structures can use a build-in notion of feature structures.

  • The left-hand side of a rule is described by a graph called a pattern; injective graph morphisms are used in the pattern matching algorithm.

  • Negative pattern can be used for a finer control on the left-hand side of rules.

  • The right-hand side or rules is described by a sequence of atomic commands that describe how the graph should be modified during the rule application.

  • Rules can be parametrized by lexical information.

  • Filters can be used at the output of each module to control the structure produced are well-formed.

  • Subset of rules are grouped in modules; the full rewriting process being a sequence of module applications.

  • The Grew software has support both for confluent and non-confluent modules; when a non-confluent modules is used, all normal forms are returned and then ambiguity is handled in a natural way.

  • Grew can be used on Corpus mode with statistics about rules usage or with an a Graphical User Interface which can show all intermediate graphs used during the rewriting process (useful either to debug rewriting system or for demonstrations).

A demonstration of the Grew Software was presented at the TALN conference in june in Grenoble.[15]

The Grew software were used for several kind of applications manipulating syntactic and/or semantic graph representations. It was used to build DMRS semantic representation from syntactic dependency trees in the French TreeBank [12] , [14] . More recently, it was used on the Sequoia TreeBank, to produce deep syntax annotation and DMRS Semantic representations.

Another application of the Grew software which is currently investigated is the detection of annotation errors in corpora. Graph Rewriting is use to detect ill-formed structures that don't fit the annotation guide requirements. In collaboration with the Alpage team, this was applied to the Sequoia Corpus and the reported errors were corrected in version 3.2 and 3.3 of the corpus(https://www.rocq.inria.fr/alpage-wiki/tiki-index.php?page=CorpusSequoia ).