Current students


LEONESIO MARCOCycle: XXXVI

Section: Systems and Control
Advisor: FAGIANO LORENZO MARIO
Tutor: PIRODDI LUIGI

Major Research topic:
Physics enhanced Artificial Intelligence for Industrial Process Control

Abstract:
The current manufacturing environment places a growing demand on autonomous control and optimization of manufacturing processes, especially in unattended machines. Nowadays, process optimization strategies based on Artificial Intelligence (AI) have not reached relevant exploitation at the machines level (except for robotic applications), despite the appropriate amount of research carried out on this topic. Indeed, pure AI strategies present some drawbacks. They need a consistent amount of data to train the system; therefore, they are not immediately effective. They ignore the treasury of apriori knowledge accumulated by long-term physics-based modeling activity and technologists' experience. Further, AI often generates distrust due to a lack of interpretability.
The proposed research seeks the integration of domain-knowledge in Machine Learning optimization techniques to: reduce the training effort; preserve the physic-based core of the control architecture in order to make the actions interpretable by experts; develop methods and interface to continuously improve control performance by a collection of sporadic expert judgments and/or new training data entries, according to a "lazy learning" approach.