[4F5] Distinguished overview speaker: Intelligent soft sensing and prediction for industrial applications
H Karimi
Politecnico di Milano, Italy
In this speech, the innovative methodologies for efficient and cost-effective process optimisation in industrial applications are explored. Soft sensing refers to approximating hard-to-measure variables using easily measurable ones at a given instant, whereas prediction extends the focus to future horizons. This presentation highlights several novel soft sensing and prediction frameworks developed for two specific case studies: industrial aluminium electrolysis and undermining systems. These frameworks leverage semi-supervised and self-supervised learning to address few-shot labelled sample scenarios while maintaining performance integrity. The talk concludes with key insights and future directions.
