yingtaoluo / Probabilistic-density-network

A Quantum Inspired Deep Probabilistic Learning Model for the Inverse Design of Meta-Structures. Research AAAS & NeurIPS ML4PS.

Probabilistic-density-network

Official Inplementation of "Probability-density-based deep learning paradigm for the fuzzy design of functional metastructures". ("A Quantum Inspired Probabilistic Learning Model for the Inverse Design of Meta-Structures" at NeurIPS 2020 ML4PS, https://arxiv.org/abs/2011.05511) A probability-density-based neural network (PDN) is proposed for the multivalued on-demand design of materials. It excels in efficiency, accuracy, and output variety than other state-of-the-art deep learning models.

Data

Link: https://pan.baidu.com/s/14YhjvUVlEV9Zq8Y4LYCyeA
Code: f16b

Key Words

Inverse Design of Metamaterials; Deep Learning Inverse Design of Materials; On-Demand Design.

Prerequisite Environment (Recommended)

For neural network training: VS2017, CUDA, cudnn, python, numpy, pytorch, sklearn, matplotlib, scipy
For numerical simulation: Matlab2017b, COMSOL5.4

Poster Presentation