Cross-Team Collaboration and Diversity in the Bridge2AI Project

Abstract

The Bridge2AI project, funded by the National Institutes of Health, involves researchers from different disciplines and backgrounds to develop well-curated AI health data and tools. Understanding cross-disciplinary and cross-organizational collaboration at the individual, team, and project levels is critical. In this paper, we matched Bridge2AI team members to the PubMed Knowledge dataset to get their health-related publications. We built the collaboration network for Bridge2AI members and all of their collaborators and sorted out researchers with the largest degree of centrality and betweenness centrality. Our finding suggests that Bridge2AI members need to strengthen internal collaborations and boost mutual understanding in this project. We also applied machine learning methods to cluster all the researchers and labeled publication topics in different clusters. Finally, by identifying the gender/racial diversity of researchers, we found that teams with higher racial diversity receive more citations, and individuals with diverse gender collaborators publish more papers.