Text Generation

Description

This book is concerned with the machine-based generation of natural language text and presents a formal analysis of problems that in the main have previously only been approached descriptively. In the process of producing discourse, speakers and writers must decide what it is they want to say and how to present it effectively. Kathleen McKeown's main concern is to identify and formalize principles of discourse so that they can be used in a computational process. The text generation theory she describes relies on principles drawn from both linguistics and computer science. It shows that by using an organizational framework for the text (a schema) and by identifying focus of attention as the text is produced, the system can produce a coherent, well-organized text. This theory has been embodied in a computer program, TEXT, which, given a question, can produce a paragraph length response. TEXT is part of a natural language database system and provides answers to general questions about the database, such as requests for definitions or queries about the kind of information available in the database. An appendix to the book provides examples of the TEXT system in operation.

' important reading for anyone either working in or just trying to keep up with natural language processing.' Computational Linguistics