No abstract

We present an implemented approach for domain-restricted question answering from structured knowledge sources, based on robust semantic analysis in a hybrid NLP system architecture. We perform question interpretation and answer extraction in an architecture that builds on a lexical-conceptual structure for question interpretation, which is interfaced with domain-specific concepts and properties in a structured knowledge base. Question interpretation involves a limited amount of domain-specific inferences, and accounts for higher-level quantificational questions. Question interpretation and answer extraction are modular components that interact in clearly defined ways. We derive so-called proto queries from the linguistic representations, which provide partial constraints for answer extraction from the underlying knowledge sources. The search queries we construct from proto queries effectively compute minimal spanning trees from the underlying knowledge sources. Our approach naturally extends to multilingual question answering, and has been developed as a prototype system for two application domains: the domain of Nobel prize winners, and the domain of Language Technology, on the basis of the large ontology underlying the information portal LT WORLD.

We address variable morphotactics, the phenomenon of order variability of morphs, in the context of inflectional morphology. Based on an extended discussion of cross-linguistic variation, including conjugation in Nepali, Fula, Swahili, Chintang and Italian, and nominal declension in Ostyak and Mari, we propose a canonical typology that identifies different deviations from strict ordering. Following a discussion of previous approaches to the problem, we propose Information-based Morphology, an inferential-realisational and model-theoretic approach to morphology couched in a logic of typed feature structures. Within this formal theory, we develop detailed analyses of the core cases in the typology and show how different types and degrees of deviation from the canon can be pin-pointed in the relative complexity of the rule type hierarchies that model the data. Furthermore, we show that complex deviations, as attested in Mari, can be understood as combinations of more basic deviations.