Methods, systems, and apparatus, including computer programs encoded on computer storage media, for reinforcement learning in the field of hygiene. Specifically, the features described relate to selecting actions in a context to be performed by an agent that interacts with an environment by receiving observations, in response, performing actions from a set of actions, wherein the context comprises a bacterial product to be prescribed, wherein the observations comprise data from the mapping of applied bacteria, and wherein the actions comprise data for the prescription of the bacterial product.