You are here

• Home
• Elsevier article award at BFAS 2024

Elsevier article award at BFAS 2024

Artificial intelligence applied to the electroerosion machining process : Loïc Guiziou1, Emmanuel Ramasso1, Sébastien Thibaud1 et Sébastien Denneulin2 won second prize for best paper at the 8th International Conference on Belief Functions.

Hosted by the University of Ulster in the heart of Northern Ireland's capital, the conference brought together international experts in Dempster-Shafer theory. Over three days, more than thirty presentations shared the latest advances in artificial intelligence, uncertainty management and fuzzy logic.

The nominated paper concerns the development of a new image clustering algorithm, aimed at automatically grouping similar data. After the training phase, this algorithm is capable of assigning new images to the corresponding group, thus learning to categorise them autonomously. A deep learning algorithm based on the theory of belief functions, it is specially designed to manage uncertainties and make decisions when faced with equivocal or noisy data. It also has the advantage of being able to operate without supervision, or to incorporate all or part of a priori, depending on requirements.

This article is part of M. Guiziou's thesis work, which focuses on the use of artificial intelligence tools applied to the electroerosion machining process - a manufacturing process that is particularly complex to model. More specifically, this algorithm is used to process acoustic data obtained in continuous flow. In the long term, this development will contribute to a better understanding of machining phenomena and will sharpen real-time decision-making operations on these machines.

1 FEMTO-ST (Département de Mécanique Appliquée, Besançon (25000)
2 SAFRAN CERAMICS, Le Haillan (33185)

Contact : Loïc Guiziou