| Summary: | In recent years, ranking approaches to Natural Language Generation have become increasingly popular. They abandon the idea of generation as a deterministic decision-making process in favour of approaches that combine overgeneration with ranking at some stage in processing. In this thesis, we investigate the use of instance-based ranking methods for surface realization in Natural Language Generation. Our approach to instance-based Natural Language Generation employs two basic components: a rule system that generates a number of realization candidates from a meaning representation and an instance-based ranker that scores the candidates according to their similarity to examples taken from a training corpus. The instance-based ranker uses information retrieval methods to rank output candidates. Our approach is corpus-based in that it uses a treebank (a subset of the Penn Treebank II containing management succession texts) in combination with manual semantic markup to automatically produce a generation grammar. Furthermore, the corpus is also used by the instance-based ranker. The semantic annotation of a test portion of the compiled subcorpus serves as input to the generator. In this thesis, we develop an efficient search technique for identifying the optimal candidate based on the A*-algorithm, detail the annotation scheme and grammar construction algorithm and show how a Rete-based production system can be used for efficient candidate generation. Furthermore, we examine the output of the generator and discuss issues like input coverage (completeness), fluency and faithfulness that are relevant to surface generation in general.
|
|---|
