Using simulation to accelerate autonomous experimentation: A case study using mechanics
Aldair E. Gongora,
Kelsey L. Snapp,
Emily Whiting,
Patrick Riley,
Kristofer G. Reyes,
Elise F. Morgan,
Keith A. Brown
Affiliations
Aldair E. Gongora
Department of Mechanical Engineering, Boston University, Boston, MA 02215, USA
Kelsey L. Snapp
Department of Mechanical Engineering, Boston University, Boston, MA 02215, USA
Emily Whiting
Department of Computer Science, Boston University, Boston, MA 02215, USA
Patrick Riley
Google Research, Mountain View, CA 94043, USA
Kristofer G. Reyes
Department of Materials Design and Innovation, University at Buffalo, Buffalo, NY 14260, USA; Corresponding author
Elise F. Morgan
Department of Mechanical Engineering, Boston University, Boston, MA 02215, USA; Department of Biomedical Engineering, Boston University, Boston, MA 02215, USA; Division of Materials Science & Engineering, Boston University, Boston, MA 02215, USA; Corresponding author
Keith A. Brown
Department of Mechanical Engineering, Boston University, Boston, MA 02215, USA; Division of Materials Science & Engineering, Boston University, Boston, MA 02215, USA; Physics Department, Boston University, Boston, MA 02215, USA; Corresponding author
Summary: Autonomous experimentation (AE) accelerates research by combining automation and machine learning to perform experiments intelligently and rapidly in a sequential fashion. While AE systems are most needed to study properties that cannot be predicted analytically or computationally, even imperfect predictions can in principle be useful. Here, we investigate whether imperfect data from simulation can accelerate AE using a case study on the mechanics of additively manufactured structures. Initially, we study resilience, a property that is well-predicted by finite element analysis (FEA), and find that FEA can be used to build a Bayesian prior and experimental data can be integrated using discrepancy modeling to reduce the number of needed experiments ten-fold. Next, we study toughness, a property not well-predicted by FEA and find that FEA can still improve learning by transforming experimental data and guiding experiment selection. These results highlight multiple ways that simulation can improve AE through transfer learning.
