Processing Unknown Words in HPSG

The lexical acquisition system presented in this paper incrementally updates linguistic properties of unknown words inferred from their surrounding context by parsing sentences with an HPSG grammar for German. We employ a gradual, information-based concept of ``unknownness'' providing a uniform treatment for the range of completely known to maximally unknown lexical entries. ``Unknown'' information is viewed as revisable information, which is either generalizable or specializable. Updating takes place after parsing, which only requires a modified lexical lookup. Revisable pieces of information are identified by grammar-specified declarations which provide access paths into the parse feature structure. The updating mechanism revises the corresponding places in the lexical feature structures iff the context actually provides new information. For revising generalizable information, type union is required. A worked-out example demonstrates the inferential capacity of our implemented system.

Automatic Inference of DATR Theories

This paper presents an approach for the automatic acquisition of linguistic knowledge from unstructured data. The acquired knowledge is represented in the lexical knowledge representation language DATR. A set of transformation rules that establish inheritance relationships and a default-inference algorithm make up the basis components of the system. Since the overall approach is not restricted to a special domain, the heuristic inference strategy uses criteria to evaluate the quality of a DATR theory, where different domains may require different criteria. The system is applied to the linguistic learning task of German noun inflection.