Institutions Involved

The network brings together six centres with considerable single-discipline expertise in different computational, theoretical and psycholinguistic aspects of dialogue research. These individual research enterprises have now reached a stage at which they can reap maximum benefit from inter-disciplinary, international collaboration.

LondonSite: King's College, London, UK (Computer Science, Philosophy): KCL has particular research strengths in foundational development of dialogue-directed grammar formalisms (Head-Driven Phrase Structure Grammar (HPSG) and Dynamic Syntax (DS)).

EdinburghSite: Universities of Edinburgh and Glasgow, UK (Human Communication Research Centre): HCRC has been at the forefront of bringing a psychological perspective on dialogue issues to psycholinguists and formal linguists, leads the challenge to the general research community that research on dialogue is central to language study, and offers expertise in empirical validation of theoretical approaches.

EssexSite: University of Essex, UK (Computer Science): Essex has expertise in theoretical and computational models of dialogue semantics (particularly for anaphora and reference, including references to the visual situation) and in annotating corpus resources to support corresponding research (some of which will be available for this project).

GothenburgSite: University of Gothenburg, Sweden (Linguistics): Gothenburg has pioneered a theoretical approach that captures higher-level aspects of dialogue, e.g. information states incorporating issues under discussion. It also has particular expertise in developing computational tools to support implementation of this approach in dialogue systems and multi-lingual grammars, including tools for extracting topic-specific sub-grammars for dialogue from broad �resource grammars�.

StonyBrookSite: State University of New York, Stony Brook, USA (Psychology, Computer Science): SUNY has unique competence in psychological investigation of processing at high and low levels of linguistic structure by both speakers and addressees, in its interaction with visual processing and eyegaze, and in combining this with computational expertise in natural language generation and dialogue management in multimodal dialogue systems.

StanfordSite: Stanford University, USA (Center for the Study of Language and Information): CSLI is a pioneer in addressing from a psycholinguistic perspective the influence of interlocutors� joint and individual goals on their dialogues, and how this relates to utterance composition and interpretation; it is also a leading centre in computational implementation of this theory in dialogue systems for human interaction with autonomous robots and intelligent software agents.

Related topics: TWikiUsers, TWikiGroups