A physics-based virtual environment for enhancing the quality of deep generative designs

Matthew Dering, James Cunningham, Raj Desai, Timothy William Simpson, Michael Andrew Yukish, Conrad S. Tucker

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper, we present a method that uses a physics-based virtual environment to evaluate the feasibility of neural network-based generated designs. Deep learning models rely on large training data sets that are used for training. These training data sets are typically validated by human designers that have a conceptual understanding of the problem being solved. However, the requirement of human training data severely constrains the size and availability of training data for computer generated models due to the manual process of either creating or labeling such data sets. Furthermore, there may be misclassification errors that result from human labeling. To mitigate these challenges, we present a physics-based simulation environment that helps users discover correlations between the form of a generated design and the physical constraints that relate to its function. We hypothesize that training data that includes machine validated designs from a physics-based virtual environment will increase the probability of generative models creating functionally-feasible design concepts. A case study involving a generative model that is trained on over 70,000 human 2D boat sketches is used to test the hypothesis. Knowledge gained from testing this hypothesis will provide human designers with insights into the importance of training data in the resulting design solutions generated by deep neural networks.

Original languageEnglish (US)
Title of host publication44th Design Automation Conference
PublisherAmerican Society of Mechanical Engineers (ASME)
ISBN (Electronic)9780791851753
DOIs
StatePublished - Jan 1 2018
EventASME 2018 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, IDETC/CIE 2018 - Quebec City, Canada
Duration: Aug 26 2018Aug 29 2018

Publication series

NameProceedings of the ASME Design Engineering Technical Conference
Volume2A-2018

Other

OtherASME 2018 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, IDETC/CIE 2018
CountryCanada
CityQuebec City
Period8/26/188/29/18

Fingerprint

Virtual Environments
Virtual reality
Physics
Labeling
Generative Models
Machine design
Boats
Neural Networks
Misclassification Error
Availability
Neural networks
Simulation Environment
Design
Training
Testing
Human
Evaluate
Requirements
Model

All Science Journal Classification (ASJC) codes

  • Mechanical Engineering
  • Computer Graphics and Computer-Aided Design
  • Computer Science Applications
  • Modeling and Simulation

Cite this

Dering, M., Cunningham, J., Desai, R., Simpson, T. W., Yukish, M. A., & Tucker, C. S. (2018). A physics-based virtual environment for enhancing the quality of deep generative designs. In 44th Design Automation Conference (Proceedings of the ASME Design Engineering Technical Conference; Vol. 2A-2018). American Society of Mechanical Engineers (ASME). https://doi.org/10.1115/DETC2018-86333
Dering, Matthew ; Cunningham, James ; Desai, Raj ; Simpson, Timothy William ; Yukish, Michael Andrew ; Tucker, Conrad S. / A physics-based virtual environment for enhancing the quality of deep generative designs. 44th Design Automation Conference. American Society of Mechanical Engineers (ASME), 2018. (Proceedings of the ASME Design Engineering Technical Conference).
@inproceedings{dacfba43cce74a5a840bee153221befe,
title = "A physics-based virtual environment for enhancing the quality of deep generative designs",
abstract = "In this paper, we present a method that uses a physics-based virtual environment to evaluate the feasibility of neural network-based generated designs. Deep learning models rely on large training data sets that are used for training. These training data sets are typically validated by human designers that have a conceptual understanding of the problem being solved. However, the requirement of human training data severely constrains the size and availability of training data for computer generated models due to the manual process of either creating or labeling such data sets. Furthermore, there may be misclassification errors that result from human labeling. To mitigate these challenges, we present a physics-based simulation environment that helps users discover correlations between the form of a generated design and the physical constraints that relate to its function. We hypothesize that training data that includes machine validated designs from a physics-based virtual environment will increase the probability of generative models creating functionally-feasible design concepts. A case study involving a generative model that is trained on over 70,000 human 2D boat sketches is used to test the hypothesis. Knowledge gained from testing this hypothesis will provide human designers with insights into the importance of training data in the resulting design solutions generated by deep neural networks.",
author = "Matthew Dering and James Cunningham and Raj Desai and Simpson, {Timothy William} and Yukish, {Michael Andrew} and Tucker, {Conrad S.}",
year = "2018",
month = "1",
day = "1",
doi = "10.1115/DETC2018-86333",
language = "English (US)",
series = "Proceedings of the ASME Design Engineering Technical Conference",
publisher = "American Society of Mechanical Engineers (ASME)",
booktitle = "44th Design Automation Conference",

}

Dering, M, Cunningham, J, Desai, R, Simpson, TW, Yukish, MA & Tucker, CS 2018, A physics-based virtual environment for enhancing the quality of deep generative designs. in 44th Design Automation Conference. Proceedings of the ASME Design Engineering Technical Conference, vol. 2A-2018, American Society of Mechanical Engineers (ASME), ASME 2018 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, IDETC/CIE 2018, Quebec City, Canada, 8/26/18. https://doi.org/10.1115/DETC2018-86333

A physics-based virtual environment for enhancing the quality of deep generative designs. / Dering, Matthew; Cunningham, James; Desai, Raj; Simpson, Timothy William; Yukish, Michael Andrew; Tucker, Conrad S.

44th Design Automation Conference. American Society of Mechanical Engineers (ASME), 2018. (Proceedings of the ASME Design Engineering Technical Conference; Vol. 2A-2018).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - A physics-based virtual environment for enhancing the quality of deep generative designs

AU - Dering, Matthew

AU - Cunningham, James

AU - Desai, Raj

AU - Simpson, Timothy William

AU - Yukish, Michael Andrew

AU - Tucker, Conrad S.

PY - 2018/1/1

Y1 - 2018/1/1

N2 - In this paper, we present a method that uses a physics-based virtual environment to evaluate the feasibility of neural network-based generated designs. Deep learning models rely on large training data sets that are used for training. These training data sets are typically validated by human designers that have a conceptual understanding of the problem being solved. However, the requirement of human training data severely constrains the size and availability of training data for computer generated models due to the manual process of either creating or labeling such data sets. Furthermore, there may be misclassification errors that result from human labeling. To mitigate these challenges, we present a physics-based simulation environment that helps users discover correlations between the form of a generated design and the physical constraints that relate to its function. We hypothesize that training data that includes machine validated designs from a physics-based virtual environment will increase the probability of generative models creating functionally-feasible design concepts. A case study involving a generative model that is trained on over 70,000 human 2D boat sketches is used to test the hypothesis. Knowledge gained from testing this hypothesis will provide human designers with insights into the importance of training data in the resulting design solutions generated by deep neural networks.

AB - In this paper, we present a method that uses a physics-based virtual environment to evaluate the feasibility of neural network-based generated designs. Deep learning models rely on large training data sets that are used for training. These training data sets are typically validated by human designers that have a conceptual understanding of the problem being solved. However, the requirement of human training data severely constrains the size and availability of training data for computer generated models due to the manual process of either creating or labeling such data sets. Furthermore, there may be misclassification errors that result from human labeling. To mitigate these challenges, we present a physics-based simulation environment that helps users discover correlations between the form of a generated design and the physical constraints that relate to its function. We hypothesize that training data that includes machine validated designs from a physics-based virtual environment will increase the probability of generative models creating functionally-feasible design concepts. A case study involving a generative model that is trained on over 70,000 human 2D boat sketches is used to test the hypothesis. Knowledge gained from testing this hypothesis will provide human designers with insights into the importance of training data in the resulting design solutions generated by deep neural networks.

UR - http://www.scopus.com/inward/record.url?scp=85057052018&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85057052018&partnerID=8YFLogxK

U2 - 10.1115/DETC2018-86333

DO - 10.1115/DETC2018-86333

M3 - Conference contribution

T3 - Proceedings of the ASME Design Engineering Technical Conference

BT - 44th Design Automation Conference

PB - American Society of Mechanical Engineers (ASME)

ER -

Dering M, Cunningham J, Desai R, Simpson TW, Yukish MA, Tucker CS. A physics-based virtual environment for enhancing the quality of deep generative designs. In 44th Design Automation Conference. American Society of Mechanical Engineers (ASME). 2018. (Proceedings of the ASME Design Engineering Technical Conference). https://doi.org/10.1115/DETC2018-86333