Dynamic interaction recommendation for student team collaboration and innovative design practice based on mixed reality technology

Abstract

In the context of educational digital transformation and the deep integration of intelligent technologies, Mixed Reality (MR) technology has enabled the creation of immersive collaborative environments, offering a novel paradigm for student team-based innovation and design practice. However, the exponential growth of interaction data and the dynamic complexity of collaborative scenarios in MR environments have presented a critical challenge: accurately capturing behavioral features within both human-computer interaction (HCI) and social collaborative interaction (SCI) to improve collaborative efficiency. Existing research demonstrates three major limitations: (i) single-modality data are insufficient to reflect the social context of team collaboration; (ii) static feature extraction methods fail to accommodate the evolving demands of design tasks; and (iii) inadequate modeling of team role differentiation has resulted in suboptimal interdisciplinary collaboration recommendations. To address these issues, a dynamic interaction recommendation method based on dual interaction relationship inference was proposed. A five-layered processing model—comprising an input layer, embedding layer, student preference extraction layer, team preference extraction layer, and prediction layer—was constructed. Dynamic student operation trajectories were captured using Gated Recurrent Unit (GRU) networks, while global and local attention mechanisms were employed to integrate team interaction data, enabling multi-level modeling of both personalized preferences and collaborative contexts. This study represents the first integration of dual interaction relationships into a recommendation framework, overcoming the limitations of conventional single-modality approaches. A dynamic, role-sensitive solution is thereby provided for intelligent collaboration systems in MR-based educational environments, offering significant theoretical and practical implications for the digital transformation of education.