Constraint-based RMRS Construction from Shallow Grammars

Constraint-based RMRS Construction from Shallow Grammars Anette Frank Language Technology Lab German Research Center for Artificial Intelligence, DFKI GmbH Stuhlsatzenhausweg 3, 66123 Saarbr¨ucken, Germany Anette.Frank@dfki.de Abstract We present a constraint-based syntax-semantics interface for the construction of RMRS (Robust Minimal Recursion Semantics) representations from shallow grammars. The architecture is de- signed to allow modular interfaces to existing shallow grammars of various depth – ranging from chunk grammars to context-free stochastic grammars. We define modular semantics con- struction principles in a typed feature structure formalism that allow flexible adaptation to al- ternative grammars and different languages.1 1 Introduction Semantic formalisms such as MRS (Copestake et al., 2003) provide elegant solutions for the treatment of semantic ambiguities in terms of underspecifi- cation – most prominently scope. In recent work Copestake (2003) has investigated a novel aspect of underspecification in the design of semantic for- malisms, which is concerned with the representation of partial semantic information, as it might be ob- tained from shallow, i.e. incomplete syntactic anal- ysis. The main rationale for this type of underspeci- fication is to ensure monotonicity, and thus upwards compatibility of the output of shallow parsing with semantic representations obtained from full syntac- tic parsing. Thus, Copestake’s design of RMRS – Robust Minimal Recursion Semantics – provides an important contribution to a novel line of research to- wards integration of shallow and deep NLP. While previous accounts (Daum et al., 2003; Frank et al., 2003) focus on shallow-deep integration at the syn- tactic level, Copestake aims at integration of shal- low and deep NLP at the level of semantics. In this paper we review the RMRS formalism de- signed by Copestake (2003) and present an archi- tecture for a principle-based syntax-semantics in- terface for RMRS construction from shallow gram- mars. We argue for a unification-based approach, 1The research reported here was conducted in the project QUETAL, funded by the German Ministry for Education and Research, BMBF, under grant no. 01 IW C02. to account for (underspecified) argument binding in languages with case-marking as opposed to structural argument identification. The architec- ture we propose is especially designed to support flexible adaptation to different types of shallow to intermediate-level syntactic grammars that may serve as a basis for RMRS construction. A chal- lenge for principle-based semantics construction from shallow grammars is the flat and sometimes non-compositional nature of the structures they typ- ically produce. We present RMRS semantics con- struction principles that can be applied to flat syn- tactic structures with various degrees of partiality. 2 RMRS – For Partial Semantic Representation Copestake (2003) presents a formalism for partial semantic representation that is derived from MRS semantics (Copestake et al., 2003). Robust Min- imal Recursion Semantics is designed to support novel forms of integrated shallow and deep NLP, by accommodating semantic representations pro- duced by NLP components of various degrees of partiality and depth of analysis – ranging from PoS taggers and NE recognisers over chunk and (non-)lexicalised context-free grammars to deep grammars like HPSG with MRS output structures. The …

Content truncated to 500 words