TY - GEN
T1 - METASET
T2 - ASME 2020 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, IDETC-CIE 2020
AU - Chan, Yu Chin
AU - Ahmed, Faez
AU - Wang, Liwei
AU - Chen, Wei
N1 - Funding Information:
We are grateful for support from the NSF CSSI program (Grant No. OAC 1835782). Yu-Chin Chan thanks the NSF Graduate Research Fellowship (Grant No. DGE-1842165). Liwei Wang would like to acknowledge support from Zhiyuan Honors Program for Graduate Students of Shanghai Jiao Tong University for his predoctoral visiting study at Northwestern University.
Publisher Copyright:
© 2020 American Society of Mechanical Engineers (ASME). All rights reserved.
PY - 2020
Y1 - 2020
N2 - Data-driven design of mechanical metamaterials is an increasingly popular method to combat costly physical simulations and immense, often intractable, geometrical design spaces. Using a precomputed dataset of unit cells, a multiscale structure can be quickly filled via combinatorial search algorithms, and machine learning models can be trained to accelerate the process. However, the dependence on data induces a unique challenge: An imbalanced dataset containing more of certain shapes or physical properties than others can be detrimental to the efficacy of the approaches and any models built on those sets. In answer, we posit that a smaller yet diverse set of unit cells leads to scalable search and unbiased learning. To select such subsets, we propose METASET, a methodology that 1) uses similarity metrics and positive semi-definite kernels to jointly measure the closeness of unit cells in both shape and property space, and 2) incorporates Determinantal Point Processes for efficient subset selection. Moreover, METASET allows the trade-off between shape and property diversity so that subsets can be tuned for various applications. Through the design of 2D metamaterials with target displacement profiles, we demonstrate that smaller, diverse subsets can indeed improve the search process as well as structural performance. We also apply METASET to eliminate inherent overlaps in a dataset of 3D unit cells created with symmetry rules, distilling it down to the most unique families. Our diverse subsets are provided publicly for use by any designer.
AB - Data-driven design of mechanical metamaterials is an increasingly popular method to combat costly physical simulations and immense, often intractable, geometrical design spaces. Using a precomputed dataset of unit cells, a multiscale structure can be quickly filled via combinatorial search algorithms, and machine learning models can be trained to accelerate the process. However, the dependence on data induces a unique challenge: An imbalanced dataset containing more of certain shapes or physical properties than others can be detrimental to the efficacy of the approaches and any models built on those sets. In answer, we posit that a smaller yet diverse set of unit cells leads to scalable search and unbiased learning. To select such subsets, we propose METASET, a methodology that 1) uses similarity metrics and positive semi-definite kernels to jointly measure the closeness of unit cells in both shape and property space, and 2) incorporates Determinantal Point Processes for efficient subset selection. Moreover, METASET allows the trade-off between shape and property diversity so that subsets can be tuned for various applications. Through the design of 2D metamaterials with target displacement profiles, we demonstrate that smaller, diverse subsets can indeed improve the search process as well as structural performance. We also apply METASET to eliminate inherent overlaps in a dataset of 3D unit cells created with symmetry rules, distilling it down to the most unique families. Our diverse subsets are provided publicly for use by any designer.
UR - http://www.scopus.com/inward/record.url?scp=85096339475&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85096339475&partnerID=8YFLogxK
U2 - 10.1115/DETC2020-22681
DO - 10.1115/DETC2020-22681
M3 - Conference contribution
AN - SCOPUS:85096339475
T3 - Proceedings of the ASME Design Engineering Technical Conference
BT - 46th Design Automation Conference (DAC)
PB - American Society of Mechanical Engineers (ASME)
Y2 - 17 August 2020 through 19 August 2020
ER -