The SEMAINE API: A Component Integration Framework for a Naturally Interacting and Emotionally Competent Embodied Conversational Agent
M. Schröder. Fachrichtung Informatik, Universität des Saarlandes, Saarbrücken, Dissertation, (January 2011)
DOI: 10.22028/D291-26253
Abstract
The present thesis addresses the topic area of Embodied Conversational Agents (ECAs) with capabilities for natural interaction with a human user and emotional competence with respect to the perception and generation of emotional expressivity. The focus is on the technological underpinnings that facilitate the implementation of a real-time system with these capabilities, built from re-usable components. The thesis comprises three main contributions. First, it describes a new component integration framework, the SEMAINE API, which makes it easy to build emotion-oriented systems from components which interact with one another using standard and pre-standard XML representations. Second, it presents a prepare-and-trigger system architecture which substantially speeds up the time to animation for system utterances that can be pre-planned. Third, it reports on the W3C Emotion Markup Language, an upcoming web standard for representing emotions in technological systems. We assess critical aspects of system performance, showing that the framework provides a good basis for implementing real-time interactive ECA systems, and illustrate by means of three examples that the SEMAINE API makes it is easy to build new emotion-oriented systems from new and existing components.
%0 Thesis
%1 Schroeder11Phd
%A Schröder, Marc
%C Saarbrücken
%D 2011
%K 01801 103 dfki book ai emotion processing middleware api multimodal user interaction agent interface zzz.mmi
%R 10.22028/D291-26253
%T The SEMAINE API: A Component Integration Framework for a Naturally Interacting and Emotionally Competent Embodied Conversational Agent
%X The present thesis addresses the topic area of Embodied Conversational Agents (ECAs) with capabilities for natural interaction with a human user and emotional competence with respect to the perception and generation of emotional expressivity. The focus is on the technological underpinnings that facilitate the implementation of a real-time system with these capabilities, built from re-usable components. The thesis comprises three main contributions. First, it describes a new component integration framework, the SEMAINE API, which makes it easy to build emotion-oriented systems from components which interact with one another using standard and pre-standard XML representations. Second, it presents a prepare-and-trigger system architecture which substantially speeds up the time to animation for system utterances that can be pre-planned. Third, it reports on the W3C Emotion Markup Language, an upcoming web standard for representing emotions in technological systems. We assess critical aspects of system performance, showing that the framework provides a good basis for implementing real-time interactive ECA systems, and illustrate by means of three examples that the SEMAINE API makes it is easy to build new emotion-oriented systems from new and existing components.
@phdthesis{Schroeder11Phd,
abstract = {The present thesis addresses the topic area of Embodied Conversational Agents (ECAs) with capabilities for natural interaction with a human user and emotional competence with respect to the perception and generation of emotional expressivity. The focus is on the technological underpinnings that facilitate the implementation of a real-time system with these capabilities, built from re-usable components. The thesis comprises three main contributions. First, it describes a new component integration framework, the SEMAINE API, which makes it easy to build emotion-oriented systems from components which interact with one another using standard and pre-standard XML representations. Second, it presents a prepare-and-trigger system architecture which substantially speeds up the time to animation for system utterances that can be pre-planned. Third, it reports on the W3C Emotion Markup Language, an upcoming web standard for representing emotions in technological systems. We assess critical aspects of system performance, showing that the framework provides a good basis for implementing real-time interactive ECA systems, and illustrate by means of three examples that the SEMAINE API makes it is easy to build new emotion-oriented systems from new and existing components.},
added-at = {2018-03-21T12:36:15.000+0100},
address = {Saarbr\"{u}cken},
author = {Schr\"{o}der, Marc},
biburl = {https://www.bibsonomy.org/bibtex/29c343d282f3c15096ae86515f722b577/flint63},
doi = {10.22028/D291-26253},
file = {SciDok SULB:2011/Schroeder11Phd.pdf:PDF},
groups = {public},
interhash = {a8da28c733c83b96cdb5ca6b2ebadbde},
intrahash = {9c343d282f3c15096ae86515f722b577},
keywords = {01801 103 dfki book ai emotion processing middleware api multimodal user interaction agent interface zzz.mmi},
month = {#jan#},
school = {Fachrichtung Informatik, Universit\"{a}t des Saarlandes},
timestamp = {2018-04-16T12:10:07.000+0200},
title = {The SEMAINE API: A Component Integration Framework for a Naturally Interacting and Emotionally Competent Embodied Conversational Agent},
type = {Dissertation},
username = {flint63},
year = 2011
}