Session: DAC-04-02: Data-Driven Design

Paper Number: 116678

116678 - When Is It Actually Worth Learning Inverse Design? 

Design optimization, and particularly adjoint-based multi-physics shape and topology optimization, is time-consuming and often requires expensive iterations to converge to desired designs. In response, researchers have developed Machine Learning (ML) approaches -- often referred to as Inverse Design methods -- to either replace or accelerate tools like Topology optimization (TO). However, these methods have their own hidden, non-trivial costs including that of data generation, training, and refinement of ML-produced designs. This begs the question: when is learning Inverse Design actually worth it, compared to just optimizing designs without ML assistance?

This paper quantitatively addresses this question by comparing the costs and benefits of three different Inverse Design ML model families on a Topology Optimization (TO) task, compared to just running the optimizer by itself.
We explore the relationship between the size of training data and the predictive power of each ML model, as well as the computational and training costs of the models and the extent to which they accelerate or hinder TO convergence. The results demonstrate that simpler models, such as KNN and RF, are more effective for TO warmstarting with limited training data, while more complex models, such as DECNN, are preferable with more data. We also emphasize the need to balance the benefits of using larger training sets with the costs of data generation when selecting the appropriate ID model. Finally, the paper addresses the challenge of using ML predictions for warmstart optimization, and provide some suggestion for budget and resource management.

Presenting Author: Milad Habibi University of Maryland, College Park

Presenting Author Biography: Milad Habibi is a current Ph.D. student in Reliability Engineering at the University of Maryland, College Park. His research is focused on scientific machine learning, optimization for engineering design, and uncertainty quantification. He received his M.Sc in Mechanical Engineering from Northern Arizona University and his B.Sc. in Mechanical Engineering from Sharif University of Technology, Iran.

Authors:

Milad Habibi University of Maryland, College Park
Jun Wang Santa Clara University
Mark Fuge University of Maryland