Deep Sentence Understanding in a Restricted Domain

Deep Sentence Understanding in a Restricted Domain* Pierre Zweigenbaum and Marc Cavazza DIAM -- INSERM U.194, 91 bd. de l'H6pital, 75634 Paris Cedex 13, France e-mail: zweig@frsim51 (bitnet) Abstract We present here the current prototype of the text understanding system HELENE. The ob- jective of this system is to achieve a deep un- derstanding of small reports dealing with a re- stricted domain. Sentence understanding builds a model of the state of the world described, through the application of several knowledge modules: (i) LFG parsing, (ii) syntactic disam- biguation based on lexical entry semantic com- ponents, (iii) assembly of semantic components and instantiation of domain entities, and (iv) construction of a world model through activa- tion of common sense and domain knowledge. 1 Introduction We present here the current prototype of the text understanding system HI~LI~NE. The ob- jective of this system is to achieve a deep un- derstanding of small reports dealing with a re- stricted domain (here, patient discharge sum- maries in a medical specialty). This means that H~LJ~NE should rely on an extensive description of all types of required knowledge. This implies of course a deep domain knowledge base, includ- ing the needed common sense knowledge. Precise understanding must not only rely on complete domain knowledge, but also on enough syntactic information. This is the reason why H~LI~NE includes a full syntactic module, whose task is to provide the semantic construction module with the (deep) structures of sentences. One problem with syntactic processing is that it gives rise to numerous ambiguities. These am- biguities are filtered on semantic grounds by a disambiguation module that does not build any semantic representation. *supported by AIM project 1003 and PRC Commu- nication Homme-Machine, PSle Langage Naturel. Semantic construction is concerned with the recognition of domain entities that can be ex- pressed by word groups. We thus had to adopt a lexical semantics approach compatible with de- scriptions. Domain entities, once instantiated, provide the basis on which a model of the cur- rent state of the world (here, the patient state) is built. The same lexical semantic in{brmation is used both to help syntactic processing, and in a more extensive way to access domain models in order to build semantic representations. The prototype includes the following main modules: The syntactic module implements the Lexi- cal Functional Grammar formalism [7]. The parser builds c-structure and f-structure bottom-up in parallel on a chart, so that f- structure validity can constrain c--structure construction. Ambiguous attachments are submitted to evaluation and ranking by the disambigua- tion module. This module applies a set of general heuristic rules that operate on the semantic definition of the LFG predicates. Semantic construction relies on dynamic domain models that integrate common sense. LFG predicates are characterized by semantic components that point to parts of the knowledge base. The prototype runs in Common Lisp (VAX Lisp) and K, a proprietary language embedded in Common Lisp. The remaining sections de- scribe these modules in more detail. 2 Parsing with a Lexical- Functional …

Content truncated to 500 words