Can AI predict 3D printing error?

Forest He, PhD student at the Visual Intelligence Laboratory, presented his recent work on manufacturability prediction using deep learning in the American Society of Mechanical Engineers (ASME) Computers and Information in Engineering (CIE) conference. ASME CIE is a premier conference in computer-aided design and manufacturing (CAD/CAM). Forest’s research has been on encoding geometric features of various objects, from CAD parts to anatomical structures in medical images, into unique, numerical “fingerprints” (i.e. vector space embedding). These fingerprints can then be correlated to various attributes such as manufacturing error, cancer survivability, etc., through machine learning techniques, leading to next-generation smarter computer systems that augment human decision makers.

Further Reading