Text Authoring, Knowledge Acquisition and Description Logics

Text Authoring, Knowledge Acquisition and Description Logics Marc Dymetman Xerox Research Centre Europe 6 chemin de Maupertuis 38240 Meylan email: marc.dymetman@xrce.xerox.com Abstract We present a principled approach to the problem of con- necting a controlled document authoring system with a knowledge base. We start by describingclosed-world au- thoring situations, in which the knowledge base is used for constraining the possible documents and orientingthe user’s selections. Then we move to open-world author- ing situations in which, additionally, choices made dur- ing authoring are echoed back to the knowledge base. In this way the information implicitly encoded in a docu- ment becomes explicit in the knowledge base and can be re-exploited for simplifying the authoring of new doc- uments. We show how a Datalog KB is sufficient for the closed-world situation, while a Description Logic KB is better-adapted to the more complex open-world situ- ation. All along, we pay special attention to logically sound solutions and to decidability issues in the different processes. Introduction Recently there has been a surge of interest in interactive natural language generation systems (Paris et al., 1995; Power and Scott, 1998; Coch and Chevreau, 2001); such systems rely on a capability of generating a natural lan- guage text from an abstract content representation, but — contrary to traditional NLG (Natural Language Gen- eration) systems — this representation is only partially available at the beginning of the text production process; it is then gradually completed by a human author, typ- ically using content-selection menus correlated with re- gions of the evolving generated text.. One such system, MDA (Multilingual Document Au- thoring) (citation omitted) is based on a formal specifi- cation — using a variant of Definite Clause Grammars (DCGs) (Pereira and Warren, 1980) — of what counts as a valid abstract content representation. The different derivation trees in the grammar correspond to texts with different contents, and at each step of the authoring pro- cess the user is asked to make interactive choices on how to expand the current partial derivation tree one step fur- ther. There are important analogies between this process and the process of authoring an XML document under the control of a DTD or a Schema, but DCGs are more expressive in terms of the contextual constraints that can be expressed and also are more adapted to the production of grammatical text.1 In published MDA work, all the knowledge about what constitutes a valid document is provided in the grammars, with no clear separation between (1) world knowledge (the fact that a certain pharmaceutical drug contains some molecule makes it dangerous for a certain patient condition) and (2) constraints about document or- ganization (if a certain drug is dangerous for a certain condition, then a warning should be generated at a cer- tain place in the document). A more principled and modular solution is to leave in the grammar all constraints pertaining to docu- ment/textual organization, and to use an external logical theory to express knowledge about the world described by the documents. A …

Content truncated to 500 words