Encoding and exploring latent design space of optimal material structures via a VAE-LSTM model

Variational autoencoders (VAE) are machine learning models that can extract low dimensional representations of data from datasets of high complexity and volume. Importantly, they can be used for generative purposes to reconstruct complex data, such as images, from a low dimensional encoding of only...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Andrew J. Lew, Markus J. Buehler
Formato: article
Lenguaje:EN
Publicado: Elsevier 2021
Materias:
T
Acceso en línea:https://doaj.org/article/7f81720314774945a003abb4058ec274
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:7f81720314774945a003abb4058ec274
record_format dspace
spelling oai:doaj.org-article:7f81720314774945a003abb4058ec2742021-12-04T04:36:02ZEncoding and exploring latent design space of optimal material structures via a VAE-LSTM model2666-359710.1016/j.finmec.2021.100054https://doaj.org/article/7f81720314774945a003abb4058ec2742021-11-01T00:00:00Zhttp://www.sciencedirect.com/science/article/pii/S2666359721000457https://doaj.org/toc/2666-3597Variational autoencoders (VAE) are machine learning models that can extract low dimensional representations of data from datasets of high complexity and volume. Importantly, they can be used for generative purposes to reconstruct complex data, such as images, from a low dimensional encoding of only a few variables. Long short-term memory (LSTM) neural networks are well suited to learning logical trajectory relationships within datasets. Using these two models in concert, we develop a VAE-LSTM approach to learn a classic mechanical materials design problem. Here, we focus on the compliance optimization of cantilever design, using a VAE to encode cantilever structures into a 2D latent space and a LSTM to learn trajectories in that latent space corresponding to the optimization process. Ultimately, we are able to clearly visualize the space of cantilever design, generate new design with extremely low density beyond the original dataset, and obtain optimal cantilever structures inspired by nature. We also demonstrate how the resulting designs can be manufactured using 3D printing, offering a rapid pathway from concept to prototype. The method we developed here can be generalized to other image-based datasets encapsulating changes from multiple factors. The ability offered by our approach to interpret complex behavior, via representations in simplified space, has great potential for application in the intelligent design and manufacturing of materials structure problems.Andrew J. LewMarkus J. BuehlerElsevierarticleMechanicsDesignMachine learning3D printingAutoencoderTopology optimizationMechanics of engineering. Applied mechanicsTA349-359TechnologyTENForces in Mechanics, Vol 5, Iss , Pp 100054- (2021)
institution DOAJ
collection DOAJ
language EN
topic Mechanics
Design
Machine learning
3D printing
Autoencoder
Topology optimization
Mechanics of engineering. Applied mechanics
TA349-359
Technology
T
spellingShingle Mechanics
Design
Machine learning
3D printing
Autoencoder
Topology optimization
Mechanics of engineering. Applied mechanics
TA349-359
Technology
T
Andrew J. Lew
Markus J. Buehler
Encoding and exploring latent design space of optimal material structures via a VAE-LSTM model
description Variational autoencoders (VAE) are machine learning models that can extract low dimensional representations of data from datasets of high complexity and volume. Importantly, they can be used for generative purposes to reconstruct complex data, such as images, from a low dimensional encoding of only a few variables. Long short-term memory (LSTM) neural networks are well suited to learning logical trajectory relationships within datasets. Using these two models in concert, we develop a VAE-LSTM approach to learn a classic mechanical materials design problem. Here, we focus on the compliance optimization of cantilever design, using a VAE to encode cantilever structures into a 2D latent space and a LSTM to learn trajectories in that latent space corresponding to the optimization process. Ultimately, we are able to clearly visualize the space of cantilever design, generate new design with extremely low density beyond the original dataset, and obtain optimal cantilever structures inspired by nature. We also demonstrate how the resulting designs can be manufactured using 3D printing, offering a rapid pathway from concept to prototype. The method we developed here can be generalized to other image-based datasets encapsulating changes from multiple factors. The ability offered by our approach to interpret complex behavior, via representations in simplified space, has great potential for application in the intelligent design and manufacturing of materials structure problems.
format article
author Andrew J. Lew
Markus J. Buehler
author_facet Andrew J. Lew
Markus J. Buehler
author_sort Andrew J. Lew
title Encoding and exploring latent design space of optimal material structures via a VAE-LSTM model
title_short Encoding and exploring latent design space of optimal material structures via a VAE-LSTM model
title_full Encoding and exploring latent design space of optimal material structures via a VAE-LSTM model
title_fullStr Encoding and exploring latent design space of optimal material structures via a VAE-LSTM model
title_full_unstemmed Encoding and exploring latent design space of optimal material structures via a VAE-LSTM model
title_sort encoding and exploring latent design space of optimal material structures via a vae-lstm model
publisher Elsevier
publishDate 2021
url https://doaj.org/article/7f81720314774945a003abb4058ec274
work_keys_str_mv AT andrewjlew encodingandexploringlatentdesignspaceofoptimalmaterialstructuresviaavaelstmmodel
AT markusjbuehler encodingandexploringlatentdesignspaceofoptimalmaterialstructuresviaavaelstmmodel
_version_ 1718372922148519936