What we do
Our basic research is in the field of distributional semantics, which captures the meaning of words or phrases in terms of the contexts in which they occur. We have developed our own approach, using data structures called APTs. Details are available here.
Our applied research looks at the application of NLP to the processing of large volumes of text documents. We have developed approaches to classification (including sentiment and attitudinal analysis), information extraction, influencer analysis, automatic tagging and automated dialogue. Much of our work is implemented in the Method52 framework (see image, left), details of which can be found here.