The use of artificial intelligence software bears the potential to shape future digital technologies by making them more personalized and attentive. This is true for a large variety of applications, but is particularly relevant in the area of health care. In order to ensure the user autonomy and control, we enable them to interact with machine learning components and personalize them to their needs.
The research project is based on the hypothesis that it is possible to improve the acceptance of such systems and further enhance their effectiveness by actively involving users in the tuning of machine learning models that form the basis of many health care applications.
The research project makes use of in-house tools that not only allow us to record behavioural data via mobile sensors, but also facilitate their analysis and interpretation. Moreover, the project will benefit from expertise in the development of social robots and virtual agents that are capable of adapting their behaviours to the user. Furthermore, there is previous work on multimodal assistive systems that tune pre-trained models on the user’s personal mobile device and thus enable privacy-preserving handling of sensitive and personal data.