A new machine learning system makes 3D printing much easier. Previously, an expert operator had to manually determine the ideal parameters that would allow efficient printing with new material, choosing the right print speed and the right amount of material.
Now an AI machine, which uses computer vision to monitor the production process and then correct errors in real-time, will be able to print objects quickly and error-free. It is even possible to create objects with special electrical or chemical properties or to make adjustments to the printing process on the fly if the material or environmental conditions change unexpectedly.
A "reinforcement learning" process was used to train the controller, in which the model is taught by trial and error with rewards. This project is really the first demonstration of building a production system that uses machine learning to train a complex control policy.