Much progress has been made in the provision of reusable data resources for Natural Language Engineering, such as grammars, lexicons, thesauruses. Although a number of projects have addressed the provision of reusable algorithmic resources (or `tools'), takeup of these resources has been relatively slow. This paper describes GATE, a General Architecture for Text Engineering, which is a freely-available system designed to help alleviate the problem. 1 Resource Reuse and Natural Language Engineering Car designers don't reinvent the wheel each time they plan a new model, but software engineers often find themselves repetitively producing roughly the same piece of software in slightly different form. The reasons for this inefficency have been extensively studied, and a number of solutions are now available (Prieto-Diaz and Freeman, 1987; Prieto-Diaz, 1993). Similarly, the Natural Language Engineering (NLE 1 ) community has identified the potential benefits of reducing repetition, and w...

An Evaluation of LOLITA and related Natural Language Processing Systems Paul Callaghan Submitted to the University of Durham for the degree of Ph.D., August 1997 --------------------- This research addresses the question, "how do we evaluate systems like LOLITA?" LOLITA is the Natural Language Processing (NLP) system under development at the University of Durham. It is intended as a platform for building NL applications. We are therefore interested in questions of evaluation for such general NLP systems. The thesis has two parts.

This paper describes a freely-available system called GATE [Cunningham, Gaizauskas, Wilks 1995] -- a General Architecture for Text Engineering -- which integrates much of this work. GATE is an

Much progress has been made in the provision of reusable data resources for Natural Language Engineering, such as grammars, lexicons, thesauruses. Although a number of projects have addressed the provision of reusable algo- rithmic resources (or 'tools'), takeup of these resources has been relatively slow.

Information Extraction Improving the ability of computer systems to process text is a significant research challenge. Many applications are based on partially structured databases, where structured data conforming to a schema is combined with free text. Information is stored as text in these applications because the queries required are not all known in advance – allowing for text is an attempt to capture information that could be relevant in the future but cannot be anticipated when the database schema is being designed. Text is also used due to the limitations of conventional databases, where the schema cannot easily be extended as new entity types and relationships arise in the future. Information Extraction (IE) is the process of finding instances of pre-defined entity types within text, while Data Integration systems build a virtual global schema from available structured data sources. We argue that combining techniques from IE and data integration is a promising approach for supporting applications that access partially structured data: the virtual global schema and associated metadata can be used to partially configure an IE process, and the information extracted by the IE process can then be integrated into the virtual global database, supporting queries which could not otherwise be answered. In this thesis we describe the design and implementation of the Experimental System To Extract Structure from Text (ESTEST) that investigates this approach. We 3 give examples of its use and experimental results from a number of application domains.