Description
TitleDesign and implementation of embodied conversational agents
Date Created2019
Other Date2019-05 (degree)
Extent1 online resource (ix, 69 pages) : illustrations
DescriptionIn less than a decade, virtual assistants had established themselves as very handy natively included components of the main commercial consumer systems available. These, expand both the applications and challenges of human-machine interaction, sparkling many solution to some well researched and new problems. Virtual agents are a manifest of how artificial intelligence is making its way into improving the regular consumer's experience, productivity, and how this, is gradually becoming an essential part of people's life. However, there is are still big gaps in those available today, in terms of the quality of the actual human-machine interaction. These agents are not fully prepared to understand the nature of actual human communication and conversation. This process not only involves understanding natural language utterances and following commands, but true conversation is achieved by also capturing information from facial expressions and body language, to finally assign a semantic meaning, in context of it all. This introduces a new layer of complexity in terms of how perception is handled, signals are processed, interleaved and analyzed, for then elicit coherent and proper behaviors and responses. This work will display the design and implementation, of a virtual modular intelligent agent. We expand the boundaries of simple unilateral communication, to a more robust, engaging, believable and meaningful interaction. This is achieved by enabling the agent to complement the basic speech input with voice emotion, facial expressions recognition, and smarter natural language parsing and generation. Finally, we demonstrate, with the presented implementation, how this agent raises the standard for human-machine communication, and how artificially intelligent agents can work in a multi-input model of complex signals in order to elicit meaningful and compelling behaviors in more natural interactions. Moreover, as opposed to other smart agents, this learns, remembers and pro-actively communicates with a complex set of coherent behaviors.
NoteM.S.
NoteIncludes bibliographical references
Genretheses, ETD graduate
LanguageEnglish
CollectionSchool of Graduate Studies Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.
Statistical Profile