Silvia Zottin

Silvia Zottin

Ph.D XXXVIII

Supervisor: Gian Luca Foresti

Room: Lab Village M15

Mail: zottin.silvia@spes.uniud.it

Research Project

Machine learning and deep learning for advanced manufacturing technologies (PNRR – Interconnected Nord-Est Innovation Ecosystem)

Currently, industry is undergoing a revolution aimed at innovation and digitization of the organizational processes of the sector. This has led to the upsurge of the advanced manufacturing technologies wherein industry equipment is largely driven by sensors and sensor data. Therefore, machine learning and deep learning techniques in the industrial field are increasingly in demand and one of the fundamental activities is anomaly detection.

Anomaly detection has become an important part of industrial automation systems, from quality control of production progress to maintaining equipment performance. So, the use of anomaly detection techniques in industrial field loads to immense positive benefits in terms of economy, time, product quality and competitiveness. The use of semi-supervised, unsupervised, or few-shot learning techniques is an effective choice to deal with the problem of imbalance in the data that characterizes anomaly detection problems and in which classical supervised approaches do not perform well.

For these reasons, the research project is aimed at the planning, development, implementation and experimental validation of models, methodologies and tools with machine learning, deep learning and data analysis techniques and their application to industrial processes of advanced manufacturing. In particular, the study and analysis of anomaly detection methods for the quality control of production lines and predictive maintenance.

This will promote the integration between university and local manufacturing industries improving their productivity and allowing them to strengthen their competitiveness on the market and be able to seize all growth opportunities.