http://www.shef.ac.uk/ http://gate.ac.uk/

 

The Department of Computer Science was founded in 1982 and since then has established national and international renown for many aspects of its teaching and research, and was recently awarded one of the top grades, Grade 5, in the nationwide 2001 Research Assessment Exercise. The Department has many international collaborations in both research and teaching; some of the companies that support its research and teaching include Daimler-Chrysler, Glaxo Welcome, Motorola and Nokia. In particular, Daimler-Chrysler has chosen the Department as one of its very few Distinguished Academic Partners.

 

The department contains leading-edge research groups in the following areas:

 

Computational Biology Computer Graphics Machine Learning Natural Language Processing Neurocomputing and Robotics Speech and Hearing Verification and Testing

 

In the new 6th Framework Programme of European research Computer Science members brought in around €3 million in grant funding for work on Information Society Technology. New projects include:

 

SEKT, Semantic Knowledge Technology, about the next generation of the Web, coordinated by British Telecom; PrestoSpace, about preserving and accessing 20th Century audiovisual media, involving the French National Archive, the BBC and Italian National TV; AMI, Augmented Multi-party Interaction, about computer enhanced multi-modal interaction in the context of meetings.

 

 

The Natural Language Processing (NLP) Group (over 50 staff) , Department of Computer Science, has focussed on robust engineering of open source software and on scientific method (quantitative evaluation and repeatability) in AI. The group has extensive experience in the fields of infrastructure for Human Language Technologies (HLT), information extraction, HLT methods for Knowledge Management and the Semantic Web, digital library support, question answering, terminology extraction, machine learning methods for HLT. They develop GATE, a General Architecture for Text Engineering1, which has many thousands of users at sites world-wide, and was recently assessed as “outstanding” and “internationally leading” by an anonymous peer review2. Other areas of research expertise include:

 

Information Extraction (IE) from the web, newswires, scientific journals and medical documents, in multiple genres and languages; Dialogue analysis and generation, and the modeling of belief structures in time; Digital Library support, e.g. in Perseus, Old Bailey and other projects; Question Answering in TREC and elsewhere; Terminology extraction and the mining of semantic relations and subsidiary information from text; The group has also been active in developing general purpose conversational engines, such as CONVERSE, a Loebner Prize winner.

 

The IE systems built using GATE have been some of the best evaluated systems in the US comparative evaluation competitions. They have been customised for a range of industries and have played a major role in the EC projects SEKT, PRESTOSPACE, KNOWLEDGEWEB, LIRICS, AMITIES, MUMIS, NAMIC, DOT.KOM, AVENTINUS and ECRAN. The shallow NLP components that come with GATE have already been adapted to a number of other languages, including most major EU languages, Bulgarian, Romanian, Chinese, Arabic, Russian, and a number of Indic languages.

In its effort to apply HLT technology in a SW context, USFD participates in several SW projects, where its research is focused on semantic annotation and automatic metadata creation. It also cooperates with several groups to promote take-up of HLT in ontology learning.

 

Recent research partnerships include some of the largest initiatives in Europe: SEKT, an Integrated Project on next generation knowledge management; PRESTOSPACE, an Integrated Project on preservation and access to cultural heritage material; KNOWLEDGEWEB, a Network of Excellence on the Semantic Web; LIRICS, an eContent project on ISO standardization of language processing resources; AKT1, a €12million project in knowledge technologies; AMITIES, an EU/DARPA project focused on dialogue processors to support multilingual human-computer interaction; MyGrid, a virtual laboratory environment for the life sciences; CLEF, a Clinical E-Science Framework; Multiflora, bioinformatics e-science for biodiversity research; FASIL, on flexible and adaptive spoken language and multi-modal interfaces; CLARITY, on cross language IR for text and audio documents; COMIC, generic cognitive models for multimodal interaction. Recently completed projects include NAMIC, on automatic authoring to link news articles in several languages, METER, measuring text reuse, MUMIS, automated multimedia metadata creation, indexing and search, and SOCIS, IE and image processing in the domain of police crime scenes.

 

 

Contacts:

 

Kalina Bontcheva - (email address)

 

Role in the project

 

In TAO USFD will be the consortium coordinator. In addition, USFD will play the leading role for research activities related to use of HLT for ontology learning and automatic augmentation of legacy content. Finally, USFD will be responsible for the first case study on providing an open-source reference showcase of the TAO methodology and infrastructure.

 

If you would like to receive news about the TAO project, you can subscribe here.