TEDx USMB talk: industrial quality control

I did my first TEDx today, TEDxUSMBAnnecy. Preparing this talk was fun. I had to adjust the plot to a mainstream audience. But keep a bit of Deep Learning for the dream of it.

I tried to merge the classical scientific discussion, with the procedural necessities of my research:

1- Raise money
2- Build a competent team
3- Experimental/practical research
4- Analysis
5- Perspectives
—repeat using your research perspectives.

The talk is on YouTube - in french, and I must caption it.
Contrôle qualité par apprentissage | Pierre Nagorny | TEDxUSMBAnnecy
The end cake is a quick demo of TheEye system.

Associated personnal bio:

After studying mechanical engineering, I completed my doctorate on non-invasive automatic control of product quality. This research subject seeks to meet the needs of modern industrial production (Industry 4.0). The objective is to guarantee perfect quality to eliminate waste and to reduce the duration of adjustments. I design methods and measurement devices that use automatic learning (Deep Learning), thanks to which humans transfer their knowledge to the machine. The machine is then able to respond to the constraints of 100% control of industrial processes: speed, repeatability, reproducibility and agility. After 3 years of Research, a startup creation with the support of Linksium and the USMB is underway.