This page lists some specific honours and masters projects I would be happy to supervise. They cover the three areas in which I carry out research:
A common way to archive legacy documents is to run them through a scanner to produce a PDF file, to which a searchable text layer is added using optical character recognition (OCR). Unfortunately, OCR is not perfect, so spelling errors are introduced that damage the effectiveness of search techniques.
Using an existing corpus of several thousand scanned academic papers (in the ACL Anthology), this project aims to develop automatic spelling correction techniques that use the corpus itself as a source of evidence for spelling corrections. For example, if the misrecognised string spe11in8 appears in a document, a simple distance metric may find other similar strings, such as spelling, to be much more frequent in the corpus, and on the basis of frequency then choose this as a correction. Of course it gets much more complicated than this, which is why i's interesting ...
Documents have a physical structure -- typically consisting of pages, columns, and paragraphs -- but they also have a logical structure, consisting of title information, sections, subsections, footnotes, tables and so on. PDF documents are primarily intended for rendering on a screen or a printer, and so are focussed on physical structure; they tend not to contain much information, if any, about the logical structure of the document. But that logical structure can be important for a variety of purposes; for example, knowing the logical structure of a document can assist in information retrieval, information extraction and text summarisation.
The aim of this project is to take a corpus of PDF documents, and to build a system that can automatically extract the logical structure of the document text, so that this can be provided in XML form for a variety of more sophisticated processing stages, or for a more flexible rendering model (for example as a hierarchically unfolding document in a web browser).
On a typical day, a researcher might receive between 5 and 10 announcements of conferences or workshops in their email. These announcements, which are typically around a printed page in length, provide important information about the event: its title, who is organising it, where it is to be held, and important dates such as the date for submissions to the conference and the deadline for registration, as well as the dates of the event itself. Some of this information is presented in a semi-structured form; other information, like the topics covered by the event, are buried in free form text.
The aim of this project is to build an information extraction system that can extract the key information from a conference announcement, and use this to populate a structured database, so that the information can be managed and searched more efficiently.
The project is suitable for someone who has a strong interest in text processing or machine learning.
The aim of this project is to build such a system with increased realism, by incorporating both appropriate facial gestures and approptiate intonation in the voice. Watch some newsreaders carefully to see how they use their facial expressions to communicate informaton, and listen to how they use prosody to increase interest in what they are saying. The challenge here is to find techniques that will allow us to derive appropriate audio visual features from a 'flat text' provided as input.
The Map Task is an experimental scenario used to gather data on how people interact in certain kinds of situations. It involves two persons . the route giver and the route follower . who are looking at their own copies of a map of an island on which there is some buried treasure. The two maps are different: the one belonging to the route giver shows the location of the treasure, and the one belonging to the route follower does not. So, the route giver has to give the route follower instructions as to how to navigate the map to find the treasure. Unfortunately, there are other differences between the two maps that mean instructions are often misunderstood or interpreted incorrectly.
The aim of this project is to build a computational agent that is able to act as either the route giver or the route follower. This involves a number of challenging subtasks: we need (a) a model of the domain using some form of knowledge representation; (b) a language generation system that can work out how a route through the map should be described; (c) a reasoning system that can work out how to recover from problems when the route follower has not understood the provided instructions; and (d) a reasoning system that can interpret the route giver's instructions to plot a path through the map.
The project is suitable for someone who has a strong interest in artificial intelligence or natural language processing.