University of Edinburgh
I work in the area of computational linguistics. We develop computer software that process
documents and speech to do some useful language-related tasks automatically.
My work focuses on tasks in the scenario of information access. Today any user has access to comprehensive
information on the World Wide Web with the aid of web search engines. But often a user's
information need does not stop with finding relevant information. He typically filters the presented
results in many ways. He rejects low quality documents, identifies more specific information from within a document or
synthesizes information from multiple documents. These tasks are time-consuming and costly to
perform on large numbers of documents, making automatic processing an attractive option.
My focus is specifically on three challenges.
- Documents on the web vary widely in quality.
- Further processing of the search results is often important, for example
summaries of the documents and translation of foreign language results.
- We want to improve access to web contents which are unlike traditional documents, for
example blogs, internet forums and online comments.
I develop techniques which provide language support for the above issues. I build systems which identity
documents with coherent and concise writing, create meaningful and
fluent summaries and translations of documents and social media text.