How Do We React to Context? Annotation of Individual and Group Engagement in a Video Corpus

One of the main challenges of recent years is to create socially intelligent machines: machines able not only to communicate but also to understand social signals and make sense of the various social contexts. In this paper we describe a new annotation method for the analysis of involvement, aiming at exploring the relation between the individual and the social context in terms of perceived involvement, starting from the idea that the group is more than the sum of its parts. We present a description of the annotation method and preliminary results of an analysis of a multi-party casual conversation extracted from a multimedia multimodal corpus. The work aims to explore the mechanisms by which we react to social context so that we can develop automatic dialogue systems that are able to adapt to their environment in a similar way.