PhD research students

Name: Mr. Darren Clowes

Start of Studies: July 2007
Country: UK
Supervisor(s): Ray Dawson and Dr Steve Probets
E-mail: D.Clowes@lboro.ac.uk
Research Group: Knowledge Management

Title: Hybrid Semantic-Document Models

In complex engineering domains, there usually exists various structured documents that an engineer is required to use and have a detailed knowledge of to perform their task. These documents normally consist of large sections of prose. It can therefore be hard to understand the precise objective of the section. These sections generally contain elements of rigour, which could be used semantically and more ambiguous prose which helps the readers understanding. One such domain is that of Tactical Data Link (TDL) standards, which are defined in the Department of Defence's MIL-STD-6016C. This document is over 7300 pages of prose and tables, with little diagrammatic information.

Using a completely rigorous approach such as a semantic model of the document can be problematic, as often the rigour obtained from the prose can be incomplete or be open to interpretation. In a complex domain such as military standards this can be quite hazardous. Similarly the current approach of using large prose has its drawbacks in that it can be incomplete as it is not clear what has been missed and there can be lots of duplication leading to inconsistencies.

As such using the TDL domain as a template, this research aims to investigate and develop a hybrid semantic-document model for documents pertaining to complex engineering domains. In particular the aims are:
* To establish that a Hybrid Semantic-Document model is both a possible and practical solution to the problem of understanding and interpreting large and complex text-based documents.
* To show that such a model and an associated toolset to use the model can bring demonstrable value to those working with the model.

The approach is expected to assist engineers through all stages of the product life cycle from design to fault tracking. This assistance is anticipated to be achieved from improved efficiency via saved time and improved understanding. Several tools will be prototyped to demonstrate the full potential benefits of this approach.