Additive manufacturing is being applied across industries, but the technology is still limited by its vulnerability to diverse errors. Expert operators are required to manually detect and correct errors and tune process variables for new parts and materials. This talk will discuss the application of recent advances in large AI models and deep learning to tackle these problems. Like humans, these AI models learn about the physics and processes behind “how a part is made”, opening the door to autonomous error detection and correction, improved part quality, and 3D printers that can learn to use new materials completely by themselves.
5 THINGS YOU WILL LEARN DURING THIS SESSION:
1. Generalisable AI models can reliably detect and range of errors and their causes.
2. If trained correctly, AI networks can enable closed-loop feedback for error correction.
3. Physics and explainable AI can be used to try and understand neural network predictions.
4. Generating large data sets correctly (and autonomously) is half the problem!
5. AI techniques can be applied across geometries, materials, printers, and setups.