Affecting People with Natural Language
Lead Research Organisation:
University of Aberdeen
Department Name: Computing Science
Abstract
Documents written in natural languages like English are increasingly being written by computers as well as people, and computer-written documents are used routinely in many specialised applications, such as weather forecasting.Unfortunately computer models of natural language generation have a relatively poor idea about how individual readers differ and which ways of saying things will be most appropriate for which readers. As a result, they are ill-suited to more challenging tasks, such as persuading a patient to give up smoking of using humour to change someone's beliefs about the significance of globall warming. These are tasks where the goal of the text is to affect the reader in a much deeper way than just to give them some important facts.The proposed research aims to improve NLG technology in terms both of sensitivity to the reader and also of the range of effects that can be achieved. For this, it is necessary for experts in the technology of NLG to learn from researchers in other fields, for instance psychology and user modelling, and for NLG researchers in turn to increase the robustness and generality of the technology to meet the new challenges.
Organisations
Publications
Edwards P
(2009)
e-Social Science and Evidence-Based Policy Assessment Challenges and Solutions
in Social Science Computer Review
Van Deemter K
(2009)
Utility and Language Generation: The Case of Vagueness
in Journal of Philosophical Logic
Thomas K
(2010)
Atlas.txt: exploring linguistic grounding techniques for communicating spatial information to blind users
in Universal Access in the Information Society
DORR B
(2010)
Interlingual annotation of parallel text corpora: a new framework for annotation and evaluation
in Natural Language Engineering
VARGES S
(2010)
Instance-based natural language generation
in Natural Language Engineering
McKinlay A
(2010)
Design issues for socially intelligent user interfaces. A discourse analysis of a data-to-text system for summarizing clinical data.
in Methods of information in medicine
Deemter V
(2010)
Not Exactly: In Praise of Vagueness
Hunter J
(2011)
BT-Nurse: computer generation of natural language shift summaries from complex heterogeneous medical data.
in Journal of the American Medical Informatics Association : JAMIA
Van Deemter K
(2012)
Toward a computational psycholinguistics of reference production.
in Topics in cognitive science
Khan IH
(2012)
Managing ambiguity in reference generation: the role of surface structure.
in Topics in cognitive science