Intelligence

We develop systems which are able to understand, structure and provide access to large, complex datasets in an efficient way. We exploit artificial intelligence, data analytics and mathematical theories to create intelligent software which utilises big data resulting from interactions between people, things and environments.

Find out about our collaborations on the Who we work with pages.

Natural language processing

Research explores ways in which computer technology can be applied to tasks which involve the use of natural (human) languages (like English, French or Arabic). Our research interests include statistical language generation, combining language and vision, text mining, emotional content and opinion in text, recommender systems, information quality, lexical representation, multilingualism, semantic metadata, evaluation methods, and natural language systems and architectures.

We apply our research in a wide range of application areas, including weather forecasts, medical information on the web, historical documents and archives, dictionaries, consumer reviews and film studies.

Computational Technologies for Visual Data

Our inter- and multi-disciplinary research includes the development and deployment of computational technologies for the capture, analysis, visualisation and fabrication of visual data, in particular 2D and 3D data. These areas include:          

  • 2D and 3D imaging techniques and workflows for recording accurate and reproducible visual datasets of physical objects or environments;
  • Extraction of semantic information of 2D and 3D shapes based on shape analysis, machine learning or computational linguistic methods;
  • Repositories for storing large visual datasets along with semantic information including provenance and legacy metadata;
  • Visualisation using web based and other interactive technologies, such as Virtual or Augmented Reality, for providing user access;
  • Design and fabrication technologies for re-using and (re)manufacturing visual data into physical objects.

Statistics and Categorical data Analysis

Our work focuses on developing regression models for correlated discrete and continuous responses, estimation methods of large covariance matrices and hypothesis testing procedures in high-dimensional settings. Although our research is motivated by cross-disciplinary collaborations with biologists, bioinformaticians, computer scientists and doctors, it finds application in a wide range of studies, including studies in social sciences, health studies and omics-studies among others. We disseminate the developed methods in R packages that are contributed to the R-CRAN and Bioconductor projects.