Abstract: Abstract. Design optimization, and particularly adjoint-based multiphysics shape and topology optimization, is time-consuming and often requires expensive iterations to converge to desired designs. In response, researchers ...

Abstract: When performing time-intensive optimization tasks, such as those in topology or shape optimization, researchers have turned to machine-learned inverse design (ID) methods—i.e., predicting the optimized geometry from input ...