Abstract

Increasing creative skills in collaborative groups is of huge interest for stakeholders in education, industry, policy making etc. However, construction of “the most” creative groups given a cohort of people and a set of common goals and tasks to perform is challenging. The complexity of this undertaking is amplified by the necessity to first understand and then measure what “the most” creative means in a particular situation. We present here our method of semi-automatic building of “the most” creative learning groups given a cohort of students and a particular learning context based on reinforcement learning (an adapted Q-learning algorithm). Various attributes that influence individual and group creativity may be considered. A case study on using this method with our Computer Science students is also included. However, the method is general and can be used for building collaborative groups in any situation, with the appropriate “the most” creative goal and attributes.

Recommended Citation

Vladoiu, M., Moise, G., & Constantinescu, Z. (2018). Towards Building Creative Collaborative Learning Groups Using Reinforcement Learning. In B. Andersson, B. Johansson, S. Carlsson, C. Barry, M. Lang, H. Linger, & C. Schneider (Eds.), Designing Digitalization (ISD2018 Proceedings). Lund, Sweden: Lund University. ISBN: 978-91-7753-876-9. http://aisel.aisnet.org/isd2014/proceedings2018/Education/9.

Paper Type

Event

Share

COinS
 

Towards Building Creative Collaborative Learning Groups Using Reinforcement Learning

Increasing creative skills in collaborative groups is of huge interest for stakeholders in education, industry, policy making etc. However, construction of “the most” creative groups given a cohort of people and a set of common goals and tasks to perform is challenging. The complexity of this undertaking is amplified by the necessity to first understand and then measure what “the most” creative means in a particular situation. We present here our method of semi-automatic building of “the most” creative learning groups given a cohort of students and a particular learning context based on reinforcement learning (an adapted Q-learning algorithm). Various attributes that influence individual and group creativity may be considered. A case study on using this method with our Computer Science students is also included. However, the method is general and can be used for building collaborative groups in any situation, with the appropriate “the most” creative goal and attributes.