Information Technology Reference
In-Depth Information
Chapter 19
Be at Odds? Deep and Hierarchical Neural
Networks for Classification and Regression
of Conflict in Speech
Raymond Brueckner and Björn Schuller
19.1
Introduction
The field of computational paralinguistics is dedicated to the study of non-verbal
elements of speech that convey information about human affect, emotion, personal-
ity, and speaker states and traits by applying mathematical models of the underlying
mechanisms, and there is an increasing amount of research in that field (Vinciarelli
et al. 2009 ; Schuller 2012 ; Schuller and Batliner 2013 ).
One important use case of computational paralinguistics is the automatic detec-
tion of conflict . Reliably detecting conflict is of high interest for the deployment of
artificial intelligence agents, customer centers, security and intelligence, and general
data mining applications, where it may be helpful to automatically find and extract
conflictual scenes in audio recordings, both on-line and off-line.
Conflict may be defined as a mode of interaction [where] the attainment of the
goal by one party precludes its attainment by the others (Judd 1978 ). The main
subjects of conflict are typically finite resources or attitude differences with respect
to an issue of interest. Conflicts might result in attempts to damage or limit the
opportunities of others, with potentially disruptive effects on the life of any group
where conflicts take place (Levine and Moreland 1998 ).
Search WWH ::




Custom Search