Visible to the public Characterizing and Modeling Linguistic Style in Dialogue for Intelligent Social Agents

TitleCharacterizing and Modeling Linguistic Style in Dialogue for Intelligent Social Agents
Publication TypeConference Paper
Year of Publication2017
AuthorsOraby, Shereen
Conference NameProceedings of the 22Nd International Conference on Intelligent User Interfaces Companion
PublisherACM
Conference LocationNew York, NY, USA
ISBN Number978-1-4503-4893-5
Keywordsargument, dialogue, Human Behavior, human factors, intelligent agents, pubcrawl, sarcasm, Scalability, Social Agents
AbstractWith increasing interest in the development of intelligent agents capable of learning, proficiently automating tasks, and gaining world knowledge, the importance of integrating the ability to converse naturally with users is more crucial now than ever before. This thesis aims to understand and characterize different aspects of social language to facilitate the development of intelligent agents that are socially aware and able to engage users to a level that was not previously possible with language generation systems. Using various machine learning algorithms and data-driven approaches to model the nuances of social language in dialogue, such as factual and emotional expression, sarcasm and humor and the related subclasses of rhetorical questions and hyperbole, we can come closer to modeling the characteristics of the social language that allows us to express emotion and knowledge, and thereby exhibit these styles in the agents we develop.
DOI10.1145/3030024.3038284
Citation Keyoraby_characterizing_2017