Finding groups through connection relationships
Relational data model with infinite flexibility
Relational data, including network graphs, such as the connection relationships of users in social networking services and the purchase histories of users for products, appear all around us. In this research, we aim to find the hidden groups in the relational data. In general, when finding groups in relational data, it is difficult to set the number and size of groups appropriately by hand. Therefore, we propose a relational data analysis method that has the ability to automatically adjust the number and size of groups in a data-driven manner according to the size and nature of the input data. By finding appropriate groups in larger relational data, we will contribute to the development of technology for more efficient information storage, search, and retrieval.
 M. Nakano, A. Kimura, T. Yamada, N. Ueda, “Baxter Permutation Process,” in Proc. Advances in Neural Information Processing Systems 33 (NeurIPS), 2020.
Masahiro Nakano / Recognition Research Group, Media Information Laboratory