Team signes

Members
Overall Objectives
Scientific Foundations
Application Domains
Software
New Results
Other Grants and Activities
Dissemination
Bibliography

Section: Software

Other On-going Software Developments

In the following, we list various on-going software developments. Most of these software and resources are in an early stage, and not yet packaged or available. Please get in touch with their correspondent to know the current status.

Hyperion: a parser for HR grammars

Participant : Richard Moot [correspondent] .

Hyperion is a hypergraph parser developed by Richard Moot. It analyses graphs using hyperedge replacement grammars. The core parser is a very small and very general implementation of Clemens Lautemann's dynamic programming algorithm for parsing graphs using hyperedge replacement grammars. A rudimentary Dot/Graphviz interface for portraying the hypergraphs and grammars is included. Hyperion is written in XSB Prolog.

Hyperion is a first implementation of many of the author's ideas which were presented at TAG+9 as well as in talks in Bordeaux and Chieti. An early version of the system can be dowloaded from http://www.labri.fr/perso/moot/hyperion/ .

Datalog parser for PMCFG

Participant : Sylvain Salvati [correspondent] .

Sylvain Salvati has developed a recognizer for PMCFG (Parallel Multiple Context Free Grammars) which has the correct prefix property. This recognizer is based on a compiler that transforms a grammar (a PMCFG) into a datalog program which in turn is optimized in several steps. Recognition is reduced to solving a query with this datalog program, the analyzed sentence being encoded as an extensional database. A datalog query solver is also provided with this prototype. The optimization of the datalog program is based on a modular approach which composes several atomic transformations. The composition of these transformations is enforcing properties of the recognition algorithm. Therefore composing different transformations yields to different recognition algorithms. This program is based on an extension of Kanazawa [50] prefix correct algorithm for MCFG (Multiple Context Free Grammars) proposed by Sylvain Salvati. This program is already used by Kanazawa and by the Calligramme INRIA project-team in Nancy. In the future it should be improved in several ways; first it should become a parser; second it should be interfaced with certain classes of Abstract Categorial Grammars, so as to perform text generation; finally, so as to face the scaling problem of handling big grammars, compositions of transformations should be compiled into one efficient transformation.


previous
next

Logo Inria