Filippo Pagani, Research Associate at the BSU, organised a postdoc away day on Monday 13th September, as part of the Bayes4Health Workshop.

The main aims of the away day were to promote collaborations between postdocs from the Bayes4Health (B4H) grant and postdocs from the MRC Biostatistics Unit, and to have a safe space to discuss research problems, brainstorm ideas, and tap into the collective expertise of the room to find new solutions and perspectives. Serendipitously, many postdocs from B4H’s sister grant CoSInES were also able to participate.

Held at the Pitt Building in Cambridge, the away day consisted of six talks given by BSU postdocs about their research, with particular emphasis on recent research and open challenges. The speakers were Sanmitra Ghosh, Pantelis Samartsidis, Bemsibom Toh, Solon Karapanagiotis, Angelos Alexopoulos, Johan van der Molen.

After each talk there was some discussion, and a break-out session was organised where attendees got together in small groups, discussing with the speakers and amongst themselves, about potential new approaches to the problems mentioned. The discussions touched on various themes including Markov chain Monte Carlo, Nonparametric statistics, Neural Networks, Variational Bayes and approximate Bayesian inference, with particular attention to the application of these methods to clinical, epidemiological, and omics data.

A total of 17 postdoc researchers attended the away day, which for most attendees was the first live in-person event since the start of the COVID-19 pandemic. The feedback from attendees was positive, and a few directly expressed their wish for a similar event to be held more often. Some of the direct impacts have been that the BSU postdocs received helpful suggestions on how to proceed with their current research, and some of the B4H and CoSInES postdocs are now participating in BSU events, including seminars and reading groups. On a very basic level, the event has helped the postdocs involved to get to know each other professionally which will make collaborations easier in the future.