The aim of research in Language, Logic and Information (LLI) is to endow computer systems with the ability to process natural language. This ability is essential for applications such as information retrieval and web search, information extraction and data mining, text summarization, and speech technology. Techniques for morphological analysis, part-of-speech tagging, word prediction, or term extraction are already in use in real-world applications in these areas, and the technology required for applications such as news summarization or spoken dialogue systems (e.g., systems that can engage in a dialogue with customers to give information about train timetables) is already at a very advanced state of development. The LLI group carries out research in semantics, corpus-based research and web applications. It covers more theoretical areas as well as very practical natural language engineering (NLE) projects.
The University of Essex is an established centre for research in computational linguistics and natural language engineering, with groups in departments including Language and Linguistics and the Data Archive as well as the School of Computer Science and Electronic Engineering (CSEE). The Language and Computation group has been created to facilitate the interaction of researchers working on computational processing of language across the University.
One of the main areas of strength of the LLI group in the School of Computer Science and Electronic Engineering is semantics and semantic processing, both from a theoretical and from an application oriented point of view. Our research in this area includes work on the logical foundations of semantics (Fox), on psychologically motivated computational models of semantic processing (Poesio), on the acquisition from texts ('text mining') of lexical and commonsense knowledge (Kruschwitz, Poesio, Sanchez-Graillet), used, e.g., in anaphora resolution (Poesio), and in applications such as web search and information retrieval (Kruschwitz, Robinson). A second area of interest is dialogue and speech. Several members of the group are interested in agent-based models of dialogue, used, e.g., in interfaces to web search systems and spoken dialogue (Fasli, Kruschwitz, Poesio), and we are actively involved in the organization of the annual workshops on the Semantics and Pragmatics of Dialogue. A third area of interest is natural language generation technology. Finally, there is a tradition at Essex for work in NL processing for the Arabic language (Poesio), and in machine translation (Arnold, Sadler).
Other researchers in the Language and Computation Group are interested in Constraint-Based grammars formalisms (Arnold, Borsley, Sadler) and in statistical and symbolic parsing (Arnold, Borsley). There is also an active interest in cognitively motivated theories of language (Clahsen, Felser, Eisenbeiss Poesio).
ARRAU (Anaphora Resolution and Underspecification). (funded by EPSRC; Poesio, in collaboration with Glasgow; 2004-2007)
Dictionaries and Bridging References (Poesio; in collaboration with Oxford University Press)
GUSTT (Guided Slicing and Targeted Transformation) (funded by EPSRC; Fox)
INSPIRE (Intelligent Support for People Orientated Process Re-engineering and Change Management). (Partially funded by the European Commission's Information Society Technologies Programme, project number IST-1999-10387.) Started in 1999 for 30 months (Fox)
TeTra (EPSRC, Fox)
GNOME (General-purpose algorithms for the generation of nominal expressions - application: web page generation). (EPSRC, Poesio, 1998-2000)
Investigating the Usefulness of Markup-Based Knowledge Extraction (funded by EPSRC; Kruschwitz, 2003-2004)
TRINDI - Information state-based models of dialogue (EU, Poesio, 1998-2000).
VENEX: An Anaphorically Annotated Corpus of Italian (Italian Government; Poesio, in collaboration with the Universita' di Venezia, 2002-2003)
YPA: Intelligent Classified Directory Enquiry Assistant (funded by BT) (Kruschwitz, 1997-2000)
The acquisition from corpora of lexical and ontological information for anaphoric processing (Poesio).
Completions and Continuations in Dialogue (Poesio, collaboration with the University of Bielefeld)
Statistical generation of referring expression for dynamic web pages (Poesio - in collaboration with Edinburgh).
Maria Fasli, Chris Fox, Udo Kruschwitz, Massimo Poesio, Jerome Robinson, Sam Steel
M-Dyaa Albakour, Jon Chamberlain,
Deirdre Lungley, Sharhida Sawani Saad, Suma Adindla, Mahmoud El-Haj
The members of the group provide PhD supervision in a range of topics.