Congratulations to Neil Newbold for successfully passing his MPhil to PhD transfer

Wednesday 11 February 2009

Abstract:

The PhD research has an overarching aim of developing a computational understanding of readability that leads naturally to a new metric. This metric is to be comprised of multiple parts, for assessing and comparing text documents. At the core of this research is the elaboration, implementation and evaluation of an 8-part framework proposed by Oakland and Lane (2004). Oakland and Lane's framework considers both textual and cognitive factors and accounts for, but does not sufficiently elaborate, considerations such as background knowledge, motivation and cognitive load as further contributors to text difficulty. The programme of research proposed aims to extend our current considerations that involves a novel integration of Natural Language Processing (NLP) techniques, demonstrated in a variety of fragmented research. Historically, readability research has focussed on measures of sentence and word length. Many have found these measures to be overly simplistic or an inadequate means to evaluate and compare the readability of documents, or even sentences.

We present work to date that has examined the limitations of current measures of readability and made consideration for how wider textual and cognitive phenomena may be accounted for. This has, to some extent, validated Oakland and Lane’s framework. We have explored how these factors are realised in the construction of prototypes that that implement several parts of this framework for: (i) document quality control for technical authors; (ii) automatic video annotation. This includes techniques for statistical and linguistic approaches to terminology extraction, approaches to lexical and grammatical simplification and known term identification.

Our consideration of readability is motivated towards making semantic content more readily available to both human readers and software systems of various kinds. Further work will address other elements of the Oakland and Lane framework and the combination of factors, for example using lexical cohesion to address issues such as syntactic development of the reader and the idea density in the text. Alongside considerations for future publications, there are already indications of support for a British Standard for readability, including an inaugural working group meeting, to which it is hoped that this work will contribute in the short-to-medium term.