Although autonomous robots can perform particularly well at highly specific tasks, learning each task in isolation is a very costly process, not only in terms of time but also in terms of hardware wearout and energy usage. Hence, robotic systems need to be able to adapt quickly to new situations in order to be useful in everyday tasks. One way to address this issue is transfer learning, which aims at reusing knowledge obtained in one situation, in a new related one. In this contribution, we develop a drumming scenario with the child robot Affetto where the environment changes such that the scene can only be observed through a mirror. In order to address such domain adaptation problems, we propose a novel transfer learning algorithm that aims at mapping data from the new domain in such a way that the original model is applicable again. We demonstrate this method on an artificial data set as well as in the robot setting.