Professor Andrea Cremaschi has been invited as a speaker at the 14th International Conference on Bayesian Nonparametrics (BNP 14), set to take place from June 23-27, 2025, at UCLA in Los Angeles, United States. This prestigious bi-annual conference brings together leading researchers and practitioners in the field of Bayesian nonparametrics, fostering discussions on cutting-edge methodologies and applications.

Andrea Cremaschi, in collaboration with B. Franzolini, will present their research “Matrix-variate priors for flexible mixture modelling of grouped data”. Their work explores novel dependent prior distributions designed to enhance Bayesian nonparametric mixture models. Unlike traditional models based on univariate species sampling processes, their approach employs a finite-dimensional matrix-variate Dirichlet distribution, allowing for greater flexibility and information sharing across groups.

This research addresses key challenges in Bayesian nonparametrics by offering a framework that supports group-specific kernels while maintaining standard prior assumptions. The proposed methodology has broad applications, ensuring interpretability and adaptability to various data structures.

For more details on Professor Cremaschi’s research, visit his academic profile on IE Research Datalab. We encourage all researchers and practitioners in the field to attend BNP 14 and take part in his talk.