Neural reparameterization improves structural optimization
arXiv:1909.04240
Abstract
Structural optimization is a popular method for designing objects such as bridge trusses, airplane wings, and optical devices. Unfortunately, the quality of solutions depends heavily on how the problem is parameterized. In this paper, we propose using the implicit bias over functions induced by neural networks to improve the parameterization of structural optimization. Rather than directly optimizing densities on a grid, we instead optimize the parameters of a neural network which outputs those densities. This reparameterization leads to different and often better solutions. On a selection of 116 structural optimization tasks, our approach produces the best design 50% more often than the best baseline method.
References in corpus (5)
- Physics-Constrained Deep Learning for High-dimensional Surrogate Modeling and Uncertainty Quantification without Labeled Data
- Global optimization of dielectric metasurfaces using a physics-driven neural network
- Freeform Diffractive Metagrating Design Based on Generative Adversarial Networks
- Topology Optimization Using Polytopes
- 3D Topology Optimization using Convolutional Neural Networks
Cited by in corpus (13)
- Generative Design by Reinforcement Learning: Enhancing the Diversity of Topology Optimization Designs
- Deep energy method in topology optimization applications
- Efficient and Modular Implicit Differentiation
- Adaptive estimation of quantum observables
- Phase Retrieval and Design with Automatic Differentiation
- ADCME: Learning Spatially-varying Physical Fields using Deep Neural Networks
- The Low-Rank Simplicity Bias in Deep Networks
- Variational Data Assimilation with a Learned Inverse Observation Operator
- Scalable Deep-Learning-Accelerated Topology Optimization for Additively Manufactured Materials
- Fast PDE-constrained optimization via self-supervised operator learning
- NSTO: Neural Synthesizing Topology Optimization for Modulated Structure Generation
- DNN-Based Topology Optimisation: Spatial Invariance and Neural Tangent Kernel
- Discovery of Physics and Characterization of Microstructure from Data with Bayesian Hidden Physics Models