The following is a tentative programme for the two days of the Symposium. Please note that it is still subject to change.
Wednesday 5th December
- 09:30 Arrivals
- 10:00 Symposium Welcome
- 10:10 Keynote 1: Justine Cassell
- 11:00 Coffee Break
- 11:30 Keynote 2: Matt Cronin
- 12:20 Lunch (for confirmed registrations)
- 13:20 Keynote 3: Jean-Marc Odobez
- 14:10 Keynote 4: Hatice Gunes
- 15:00 Coffee Break
- 15:30 Plenary brainstorm about possible cluster topics of interest.
- 16:15 Round Table Breakout Discussions
- 17:00 Plenary Report Back
- 17:45 Borrel and Drinks
- 18:45 Leave for Dinner (at your own cost)
Thursday 6th December
- 09:00 Keynote 5: Steve Kozlowski
- 09:50 Keynote 6: Michaela Kolbe
- 10:40 Coffee Break (Group Photo Moment)
- 11:10 Plenary Brainstorm for more discussion topics (after a night of reflection)
- 11:40 Round Table Breakout Discussions
- 12:00 Lunch
- 13:00 Round Table Breakout Discussions
- 14:00 Plenary Report Back and determining topics for final break out sessions
- 14:45 Round Table Breakout Discussions
- 15:30 Coffee Break
- 16:00 Plenary Report Back, Action Plans, and Closing.
Finishes at 5pm.
Keynote 1: Curiosity: a group phenomenon? – Justine Cassell
As a way of launching the discussion, I’m going to discuss units of analysis in research on groups – do we see groups as one beast with many arms, or a bunch of nodes connected to one another (or any one of a number of other paradigms)? I believe the answers change the way we analyze group behavior – and our results – in fundamental ways. These questions will be exemplified by discussing research my group and I have been carrying out on the social nature of curiosity – specifically, whether and how curiosity may be evoked by others rather than and in addition to one’s own prior behavior, and what it means for an individual to be curious within a group as opposed to a group being curious, and so forth.
Keynote 2: Representational Gaps: Threats to and Opportunities for Interdisciplinary Work – Matthew Cronin
Often the senders and receivers of scientific communication have different knowledge bases. While such communication is essential for solving the complex social and technological problems that affect multiple stakeholders, a diversity of knowledge among communicators can create representational gaps (rGaps). RGaps occur when senders make assumptions that receivers do not, creating conflict over the meaning and value of the information communicated. Such conflict could, if managed, promote learning and innovation as communicators reconcile their assumptions. More often, however, rGaps cause conflict to transform from a debate that informs to an argument that divides. Managing rGap conflict so that it does not degrade communication requires relationship building to mitigate the negative by-products of persistent conflict while maintaining appropriate levels of cognitive distinctiveness among diverse stakeholders. Thus, we provide a framework for identifying and leveraging rGaps through managed conflict so that communication between those with different perspectives builds rather than burns bridges.
Keynote 3: Robust perception for human behavior analysis – Jean-Marc Odobez
In the future, more and more robots or systems in general will co-exist with humans in dynamic environments. In order for them to understand the social scene, monitor or pro-actively interact autonomously and naturally with people or groups of people, it is crucial to endow them with efficient multi-modal sensing and situated perception capabilities. At the first level, these sensing methodologies should be able to analyse the different streams of information (audio, vision, depth, robot or system state, context) to detect and track the perceived state of people along several dimensions: physical state (location, trajectory, pose, speaking status), communication or social states (engagement, addressee, floor control; age, gender, rapport), which both can rely on non-verbal cues related to body language, activity state, etc. This information could then be further used to infer higher level states like mood, personality, and state of mind in general. In this talk, I will briefly present different works related to the first level of analysis, revolving around sound and human voice localization, 3D head pose tracking under 360 degrees, as well as gaze, attention, or head gesture recognition. Illustrations on how these cues have been exploited for behavior analysis will be provided as well. In general, our main goals were to investigate the use of computer vision and deep neural network architectures for the different sensing tasks, including relying on synthetic data for obtaining sufficient amount of training data, building personalized models through online learning or adaptation, and potentially taking advantage of priors on social interactions to obtain weak labels for model adaptation.
Keynote 4: Affect and Personality Analysis in Multi-person Settings. – Hatice Gunes
Computing that is socio-emotionally intelligent aims to equip devices, interfaces and robots with the means to interpret, understand, and even respond and adapt to human nonverbal behaviour, personality, affect, moods and intentions, similarly to how humans rely on their senses to assess each other’s affective and social behaviour. Creating systems and interfaces with socio-emotional intelligence is not easy. Past works have mainly focussed on automatically analysing expressions, affect and personality of people in individual settings. Affect and personality analysis in multiperson settings is challenging as it calls for new definitions, new datasets with meaningful annotations, and appropriate feature extraction and classification mechanisms in space and time. This talk will present an overview of some of the works we have conducted in this area in the context of human-media interaction and human-robot interaction.
Keynote 5: Capturing Team Dynamics: Team Interaction Sensors – Steve Kozlowski
Digital traces (e.g., email, web scraping, etc.) and new technologies (e.g., activity monitors) have social science poised at the forefront of a revolution in which we may be able to measure actual ongoing behavior rather than relying on retrospective perceptions of behavior. These technologies have unique features and challenges that have to be resolved to be applied effectively in research. My research teams are focused on capturing team process dynamics. We investigate team dynamics using video (e.g., medical trauma teams), computational modeling / agent-based simulation (e.g., team knowledge emergence; multi-team systems and adaptation), and a new technology under development – team interaction sensors (aka sociometric badges) – that are the focus of this presentation. The team interaction sensor (TIS) system we have developed at Michigan State University (MSU) with NASA support (i.e., “MSU badge”) streams dyadic interaction and physiological data in real time. My presentation will focus on describing the features, development, validation, and research findings using the MSU badge to examine team functioning in NASA mission simulations (HERA and HI-SEAS) that emulate the isolated, confined, and extreme (ICE) conditions of long duration space missions as part of our NASA research.
Keynote 6: Measuring Team Dynamics – Challenges and Needs from a Simulation-Based Training Perspective – Michaela Kolbe
I will talk about requirements of team dynamics research for medical simulation-based team training (SBTT). Improving patient safety by training teams to successfully manage emergencies is a major concern in healthcare. Most current trainings use simulation of emergency situations to practice and reflect on relevant clinical and behavioural skills. Team research, in particular team dynamics research, could offer important content and inform SBTT design and conduct. However, many clinical simulation educators lack the access to applicable evidence from team science. Instead, they have to rely on superficial knowledge about team functioning which stands in contrast to their in-depth clinical knowledge. As a result, clinical aspects are discussed with more emphasis and evidence than team aspects, leaving the potential of SBTT not fully exploited. I will address the respective challenges and needs and highlight what kind of team dynamics research findings would be beneficial for improving SBTT, and ultimately patient safety.