Motivations

Human face-to-face communication is a little like a dance, in that participants continuously adjust their behaviors based on verbal and nonverbal displays and signals. A topic of central interest in modeling such behaviors is the patterning of interlocutor actions and interactions, moment-by-moment, and one of the key challenges is identifying the patterns that best predict specific actions. Thus we are interested in developing predictive models of communication dynamics, able to integrate previous and current actions from all interlocutors to anticipate the most likely next actions, of one or all interlocutors. Humans are good at this: they have an amazing ability to predict, at a micro-level, the actions of an interlocutor; and we know that better predictions can correlate with more empathy and better outcomes.

With turn-taking being perhaps the best-known example, we now know a fair amount about some aspects of communication dynamics, but much less about others. However, recent advances in machine learning and experimental methods, and recent findings from a variety of perspectives, including conversation analysis, social signal processing, adaptation, corpus analysis and modeling, perceptual experiments, and dialog systems-building and experimentation, mean that the time is ripe to start working towards more comprehensive predictive models.

Building predictive models of human communication will enable:

• Better interactive systems (robots, embodied conversational agents, intelligent virtual agents, spoken dialog systems etc.),
• More accurate understanding and indexing of recorded dialogs and meetings,
• The ability to describe and teach cultural norms,
• The ability to diagnose mismatches in behavior and expectation that hurt effective communication in the workplace and in social relationships,
• Deeper understanding of various communication disorders,
• Better understanding of human nature.

The goal of this workshop is to bring researchers with diverse backgrounds and perspectives together to share knowledge, discuss directions, and perhaps lay the seeds for the development of a community of interest. In particular we will also be looking for answers to three questions:

• What are the most important research questions in this area?
• What obstacles are currently limiting progress?
• How can we work together in the future?