Exploiting the Interplay between Social and Task Dimensions of Cohesion to Predict its Dynamics Leveraging Social Sciences

Abstract

Emergent states are behavioral, cognitive and affective processes appearing among the members of a group when they interact together. In the last decade, the development of computational approaches received a growing interest in building Human-Centered systems. Such a development is particularly difficult because some of these states have several dimensions interplaying somehow and somewhere over time. In this paper, we focus on cohesion, its dimensions and their interplay. Several definitions of cohesion exist, it can be simply defined as the tendency of a group to stick together to pursue goals and/or affective needs. This plethora of definitions resulted in many different cohesion dimensions. Social and Task dimensions are the most investigated both in Social Sciences and Computer Science since they both play an important role in a wide range of contexts and groups. To the best of our knowledge, however, no previous work on the prediction of cohesion dynamics focused on how these 2 dimensions interplay. We leverage Social Sciences to address this issue. In particular, we take advantage of the importance of Social cohesion for creating flexible and constructive relationships to reinforce Task cohesion. We describe a Deep Neural Network architecture (DNN) for predicting the dynamics of Task cohesion by applying transfer learning from a pre-trained model dedicated to the prediction of Social cohesion dynamics. Our architecture is evaluated against several baselines. Results show that it significantly improves the predictions of the Task cohesion dynamics, confirming the benefits of integrating Social Sciences insights into models architectures. CCS CONCEPTS• Human-centered computing → Collaborative and social computing; • Computing methodologies → Artificial intelligence.

Deep Learning Analysis of Surgical Video Recordings to Assess Nontechnical Skills

“…15,16 Consequently, a shift toward automated and real-time research methods is imperative to adequately reflect the nuanced and dynamic nature of team processes and effectiveness in surgical environments. [17][18][19] Evidence from sports science and team analysis 11 shows the significant association between team members' physical movements and overall team performance, 20,21 including aspects of communication, coordination, and situational awareness. For instance, motion analysis in sport science has been effectively used to predict team success, highlighting patterns of movement that correlate with strategic teamwork and performance outcomes.…”

Few Labels are Enough! Semi-supervised Graph Learning for Social Interaction

“…For example, Hung et al classified teams as high or low on cohesion from manually annotated meetings recordings [26]. Maman and colleagues [37] analyzed the temporal dynamics of cohesion by using the GAME-ON dataset [36], which was recorded and manually annotated for the purpose. Lee et al [32] developed a computational model of interpersonal trust through hand-coded nonverbal social behaviors and explored the temporal dynamic of the construct by means of hidden Markov models.…”

“…Since feature extraction is not the focus of this research, we leverage the state-of-the-art on analysis of team behavior and emergent states (e.g., [22,37,13,15]). We select a collection of features that were already successfully used in existing works.…”

“…Specifically, we use OpenSmile [17] to compute the functional set of the Geneva Minimalistic Acoustic Parameter Set (GeMAPS) [16]. Such a set was previously successfully exploited to classify cohesion dynamics [37] and for predicting team performances [57]. Since GeMAPS loudness can be affected by the recording settings of the specific dataset, we normalize its mean and standard deviation with respect to those computed on the whole source audio.…”

“…Regarding movement features, we extract two individual features, i.e., Upper Body Energy (UBE) [13,37] and the distance of the head of each team member from the team barycenter (BarDist) [22,13,37]. We also extract one feature concerning the team as one, i.e., the entropy of the displacement of the heads (HeadsEnt) [15].…”