Artificial intelligence helps design an ultra-aerodynamic bike

Our program is based on a machine-learning system that learns to develop an intuition about the laws of physics. It can then optimize the shape to get the best possible performance. In this case, we are provided by engineers with constraints that have to fit inside the shell for an aerodynamic bike. The standard machine-learning algorithm we use to work with in our lab takes images as input. An image is a very well-structured signal that is very easy to handle by a machine-learning algorithm. However, for engineers working in this domain, they use what we call a “mesh”. A mesh is a very large graph with a lot of nods that is not very convenient to handle for a standard machine-learning algorithm. What we had to do is to develop a machine-learning algorithm that is able to handle this kind of input and learn the laws of physics from this input. Neural Concept is an EPFL spin-off that commercializes this technology. Indeed, it can also be applied to optimize drones, windmills, airplanes, or automobiles. But this technology can also be applied to optimize other computational fluid dynamics problems


Add a Comment

Your email address will not be published. Required fields are marked *