project-26-baybes.md

number	title	topic	team_leads	contributors	github	youtube_video
26	Multiple-Context Bayesian Optimization	general	Joscha Hoche (Merck KGaA / EMD group) @hochej Viola Muning Li (Merck KGaA / EMD group) @vola-m-li Marcel Mueller (Merck KGaA / EMD group and University of Bonn) @marcelmbn, @marcelmuellergdi Rim Rihana (Merck KGaA / EMD group) @RimRihana Tobias Ploetz (Merck KGaA / EMD group) @tobiasploetz	Martin Fitzner (Merck KGaA / EMD group) @Scienfitz Alexander Hopp (Merck KGaA / EMD group) @AVHopp	AC-BO-Hackathon/project-26-multiple-context-bo	wK266A0TvZ4

Traditionally, Bayesian Optimization (BO) is performed for a specific optimization task, e.g., for optimizing a cell culture medium for a specific cell type. If the medium is to be optimized for a different cell type, a new, uncorrelated optimization campaign is started. In multi-context BO (which could also be referred to as transfer learning), the information already available about the medium optimization campaign for the previous cell type is inherited into the new run.

In this project, we aim to investigate (i) the multiple-context performance of BO frameworks on existing benchmarks, (ii) develop new benchmarks for such tasks based on existing data, and (iii) possibly also investigate different ways of incorporating prior knowledge into the model.

Check our submission post here on X and here on LinkedIn!

References:

• https://github.com/emdgroup/baybe
• Transfer Learning in BayBE