You are hereMotivational scenario for collective adaptation in hybrid populations

Motivational scenario for collective adaptation in hybrid populations


Our motivational scenario for studying collective adaptation in hybrid populations concerns a setting where humans are surrounded by devices that can perform all kinds of observations about human behavior and functioning and can support them in a highly personalized way by performing certain actions or providing advice. Here, there can be a variety of devices that contain different sensors to perform observations and actuators to provide support, think of smart phones, smart watches, sensors in smart homes, etcetera. Essentially, we are aiming to generate good controllers for these devices that map sensory inputs to actions. There are a number of challenges that underlie this scenario as shown in the figure below:


  1. How can we learn from all the experiences gathered from the various devices across multiple people, can we for instance extract meaningful generic predictive models on the effectiveness of certain actions and the course of human behavior over time?

  2. How can the generic models be utilized in each individual device? How can the models be personalized towards an individual user, and how can the controller be adapted based on user feedback?

  3. Devices should learn how to work together in a proper way, when should they share certain sensory information? And how can they learn from each other’s controllers?

Finding appropriate techniques to tackle the challenges sketched above is far from trivial and require a number of fundamental issues to be solved. Our group is striving to develop algorithms to try and solve each of the issues above thereby minimizing the burden for users.

Example publications of students and CI group members:


  1. Kop, R., Hoogendoorn, M., and Klein, M.C.A., A Personalized Support Agent for Depressed Patients: Forecasting Patient Behavior Using a Mood and Coping Model. In: Slezak, D., Dunin-Keplicz, B., Lewis, M., and Terano, T. (eds.), Proceedings of the 2014 IEEE/WIC/ACM International Conference on Intelligent Agent Technology (IAT 2014), IEEE Computer Society Press, 2014, pp. 302-309.

  2. Hoogendoorn, M., Moons, L.G., Numans, M., and Sips, R.J., Utilizing Data Mining for Predictive Modeling of Colorectal Cancer using Electronic Medical Records. In: Proceedings of the 2014 Brain Informatics and Health Conference, Lecture Notes in Computer Science vol. 8609, 2014, pp. 132-141.

  3. Hoogendoorn, M., Predicting Human Behavior in Crowds: Cognitive Modeling versus Neural Networks. In: Ali, M. et al. (eds.), Recent Trends in Applied Artificial Intelligence, Proceedings of the 26th International Conference on Industrial, Engineering & Other Applications of Applied Intelligent Systems, IEA-AIE 2013, Springer, LNCS 7906, 2013, pp. 73-82.

  4. Ven, P. v.d., Henriques, M.R., Hoogendoorn, M., Klein, M., McGovern, E.,Nelson, J., Silva, H., and Tousset, E., A Mobile System for Treatment of Depression. In: Proceedings of the 2nd International Workshop on Computing Paradigms for Mental Health (MindCare 2012), 2012, pp. 47-58.

  5. Both, F., and Hoogendoorn, M., Utilization of a Virtual Patient Model to Enable Tailored Therapy for Depressed Patients. In: Lu, B.L., Zhang, L., and Kwok, J., Neural Information Processing, 18th International Conference, ICONIP 2011. LNCS 7064, Springer, 2011, pp. 700-710.