Towards supporting shared cognition in distributed multidisciplinary teams: design and evaluation of an AI-based affective recognition system for online collaboration

Abstract

This thesis investigates the challenges and opportunities of multidisciplinary design collaboration in distributed teams, with a particular focus on creativity and shared under- standing. Using a Design Research Methodology (DRM) structure, this thesis leverages a literature review and qualitative research for the Descriptive Study, proposes a prototype for the Prescriptive Study, and provides an evaluation of the tool for the Descriptive study to help multidisciplinary teams improve distributed collaboration based on rigorous guidance on how to do so in practice: 1. What is our current understanding of multidisciplinary collaboration in relation to design teamwork? 2. How have previous studies been conducted through their experimental choices, and adopted setups? 3. How do we build a shared cognition with our teammates in distributed design collaboration? 4. How does affective recognition (including sentiment and emotion) affect shared understanding and creativity in text-based communication?

To answer these, we conducted a systematic literature review that identifies the main patterns and discrepancies in previous studies on multidisciplinary design collaboration, followed by an interview-based qualitative study that investigates the use of virtual collaboration tools and their impact on distributed collaboration and shared cognition. We developed a prototype with a machine learning-based dynamic affective recognition feed-back system and conducted mixed-methods evaluation studies of the tool. Research shows that the prototype was effective in increasing both creativity and shared understanding in distributed multidisciplinary design collaboration.

The work draws on the research at the intersection of design studies, human-computer interaction, and team management, providing a better understanding of multidisciplinary distributed collaboration and suggests directions for AI solution designers looking to augment decentralised teamwork. Key contributions of this research include findings from a systematic literature review of prior empirical studies, insights from a qualitative study for improving the use of virtual tools in multidisciplinary design, an evidence-based prototype with an affective recognition feedback system which increases shared understanding and creativity in distributed design collaboration, and a set of guidelines for designing future AI-based tools.