Range-GAN: Range-Constrained Generative Adversarial Network for Conditioned Design Synthesis (2021-03-10T00:00:00.000000Z)

TL;DR

A conditional deep generative model, Range-GAN, is proposed to achieve automatic design synthesis subject to range constraints and addresses the sparse conditioning issue in data-driven inverse design problems by introducing a label-aware self-augmentation approach.

Abstract

Typical engineering design tasks require the effort to modify designs iteratively until they meet certain constraints, i.e., performance or attribute requirements. Past work has proposed ways to solve the inverse design problem, where desired designs are directly generated from specified requirements, thus avoid the trial and error process. Among those approaches, the conditional deep generative model shows great potential since 1) it works for complex high-dimensional designs and 2) it can generate multiple alternative designs given any condition. In this work, we propose a conditional deep generative model, Range-GAN, to achieve automatic design synthesis subject to range constraints. The proposed model addresses the sparse conditioning issue in data-driven inverse design problems by introducing a label-aware self-augmentation approach. We also propose a new uniformity loss to ensure generated designs evenly cover the given requirement range. Through a real-world example of constrained 3D shape generation, we show that the label-aware self-augmentation leads to an average improvement of 14% on the constraint satisfaction for generated 3D shapes, and the uniformity loss leads to a 125% average increase on the uniformity of generated shapes’ attributes. This work laid the foundation for data-driven inverse design problems where we consider range constraints and there are sparse regions in the condition space. For further information and code for this paper please refer to http://decode.mit.edu/projects/rangegan/.

Authors

Faez Ahmed

5 papers

A. Nobari

3 papers

Wei Chen

2 papers

TL;DR

Abstract

Authors

References39 items

Copyright 2021

IH-GAN: A Conditional Generative Model for Implicit Surface-Based Inverse Design of Cellular Structures

PaDGAN: Learning to Generate High-Quality Novel Designs

Deep Generative Model for Efficient 3D Airfoil Parameterization and Generation

CcGAN: Continuous Conditional Generative Adversarial Networks for Image Generation

3D Design Using Generative Adversarial Networks and Physics-Based Validation

Deep Generative Modeling for Mechanistic-based Learning and Design of Metamaterial Systems

Airfoil Design Parameterization and Optimization using Bézier Generative Adversarial Networks

Conditional Generative Adversarial Network Framework for Airfoil Inverse Design

Development of a Conditional Generative Adversarial Network for Airfoil Shape Optimization

A Case Study of Deep Reinforcement Learning for Engineering Design: Application to Microfluidic Devices for Flow Sculpting

3D Shape Synthesis for Conceptual Design and Optimization Using Variational Autoencoders

Deep Generative Design: Integration of Topology Optimization and Generative Models

Probabilistic Representation and Inverse Design of Metamaterials Based on a Deep Generative Model with Semi‐Supervised Learning Strategy

DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation

Occupancy Networks: Learning 3D Reconstruction in Function Space

Learning Implicit Fields for Generative Shape Modeling

Global-to-local generative model for 3D shapes

Kinematic Synthesis Using Reinforcement Learning

Multi-chart generative surface modeling

Microstructural Materials Design Via Deep Adversarial Learning Methodology

Nanophotonic particle simulation and inverse design using artificial neural networks

Learning Representations and Generative Models for 3D Point Clouds

Modulating early visual processing by language

Synthesizing 3D Shapes via Modeling Multi-view Depth Maps and Silhouettes with Deep Generative Networks

Self-Normalizing Neural Networks

Learning a Hierarchical Latent-Variable Model of 3D Shapes

Conditional Image Synthesis with Auxiliary Classifier GANs

Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling

Estimating and Exploring the Product Form Design Space Using Deep Generative Models

ShapeNet: An Information-Rich 3D Model Repository

Learning Structured Output Representation using Deep Conditional Generative Models

Conditional Generative Adversarial Nets

Reinforcement learning of motor skills in high dimensions: A path integral approach

Marching cubes: A high resolution 3D surface construction algorithm

AUTO-ENCODING VARIATIONAL BAYES

Circuit Synthesis Using Generative Adversarial Networks (GANs)

GENERATIVE ADVERSARIAL NETS

We propose label-aware self-augmentation, which allows for augmenting the data such that the sparsely populated regions of the condition (label) space are populated

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names