Paper Type
Complete
Paper Number
PACIS2025-1398
Description
Scientific collaboration is still largely shaped by explicit connections, such as citation networks and incidental factors like chance meetings at conferences. This over-reliance on conventional yet limited networks constrains the discovery of conceptually related but unlinked research. To address this issue, we propose a novel approach that integrates citation analysis and semantic embeddings to uncover latent collaboration opportunities. By contrasting explicit (citation-based) and latent (embedding-based) connections, we identify research areas where semantic similarity is high despite weak citation ties, revealing overlooked interdisciplinary potential. A case study at Kyoto University demonstrates how this method highlights meaningful but previously unrecognized research overlaps. Interviews with researchers further confirm that embedding-based insights can reveal genuine intellectual connections that traditional methods fail to capture. By systematically bridging explicit and latent research ties, our approach enables a more deliberate, scalable, and data-driven strategy for fostering interdisciplinary collaboration, moving beyond the constraints of existing academic networks.
Recommended Citation
Onozuka, Ryo; Shimpei, Akita; Okumura, Hideyuki; and Ogawa, Takaya, "Revealing Hidden Interdisciplinary Collaborations: A Novel Approach Using Citation Analysis and Semantic Embeddings" (2025). PACIS 2025 Proceedings. 15.
https://aisel.aisnet.org/pacis2025/aiandml/aiandml/15
Revealing Hidden Interdisciplinary Collaborations: A Novel Approach Using Citation Analysis and Semantic Embeddings
Scientific collaboration is still largely shaped by explicit connections, such as citation networks and incidental factors like chance meetings at conferences. This over-reliance on conventional yet limited networks constrains the discovery of conceptually related but unlinked research. To address this issue, we propose a novel approach that integrates citation analysis and semantic embeddings to uncover latent collaboration opportunities. By contrasting explicit (citation-based) and latent (embedding-based) connections, we identify research areas where semantic similarity is high despite weak citation ties, revealing overlooked interdisciplinary potential. A case study at Kyoto University demonstrates how this method highlights meaningful but previously unrecognized research overlaps. Interviews with researchers further confirm that embedding-based insights can reveal genuine intellectual connections that traditional methods fail to capture. By systematically bridging explicit and latent research ties, our approach enables a more deliberate, scalable, and data-driven strategy for fostering interdisciplinary collaboration, moving beyond the constraints of existing academic networks.
Comments
AI ML