Study and development of reinforcement learning techniques for smart manufacturing
The research programme aims at studying and developing reinforcement learning techniques for smart manufacturing. The creation of smart factories requires that industrial machines/robots along the production line are able to change their own control parameters in order to either prevent faults or restore the normal functioning of the production line. The reinforcement learning techniques are a tool that may allow to achieve such goal, but they need to be adapted to domains where the parameter exploration is not free since it is risky. Hence, the goal of this research is the definition of learning algorithms that are both effective and safe.
This job comes from a partnership with Science Magazine and