Inverse design of two-dimensional graphene/h-BN hybrids by a regressional and conditional GAN
Inverse design of two-dimensional graphene/h-BN hybrids by a regressional and conditional GAN
复制标题
DOI：
10.1016/j.carbon.2020.07.013
复制
发表时间：
2020-11-01
期刊：
影响因子：
10.9
通讯作者：
Lin, Jian
中科院分区：
文献类型：
--
作者：
Dong, Yuan;Li, Dawei;Lin, Jian
设计具有所需特性的材料目前是费力的，并且严重依赖于研究人员通过试错过程的直觉。为了应对这一挑战，我们提出了一种新的回归和条件生成对抗网络（RCGAN），用于逆向设计代表性的二维材料，石墨烯和氮化硼（BN）混合物。RCGAN结合了一个有监督的回归网络，从而克服了传统无监督GAN中的常见技术障碍，即在输入连续和定量标签时无法生成数据。RCGAN可以自主生成石墨烯/BN混合物，给出任何目标带隙值。这些结构与用于训练的结构不同，并且对于给定的带隙表现出高多样性。此外，它们表现出高保真度，产生的带隙在所需带隙的10%MAE（F）内，如密度泛函理论（DFT）计算所验证的。通过主成分分析（PCA）和改进的局部线性嵌入（MLLE）分析表明，该生成器成功地生成了符合真实的结构统计分布的结构。这意味着RCGAN在识别隐藏在高维数据中的物理规律的可能性。设计回归GAN架构的新策略以及在材料逆向设计中的成功应用将激发材料以外研究领域的进一步探索。(C)2020爱思唯尔有限公司版权所有。
Design of materials with desired properties is currently laborious and heavily relies on intuition of researchers through a trial-and-error process. To tackle this challenge, we propose a novel regressional and conditional generative adversarial network (RCGAN) for inverse design of representative two-dimensional materials, the graphene and boron-nitride (BN) hybrids. RCGAN incorporates a supervised regressor network, thus overcoming the common technical barrier in the traditional unsupervised GANs, which cannot generate data when fed with continuous and quantitative labels. RCGAN can autonomously generate graphene/BN hybrids given any target bandgap values. These structures are distinguished from the ones used for training and exhibit high diversity for a given bandgap. Moreover, they exhibit high fidelity, yielding bandgaps within similar to 10% MAE(F) of the desired bandgaps as validated by density functional theory (DFT) calculations. Analysis by the principle component analysis (PCA) and modified locally linear embedding (MLLE) reveals that the generator has successfully generated structures following the statistical distribution of the real structures. It implies the possibility of the RCGAN in recognizing physical rules hidden in the high-dimensional data. The novel strategy for designing regressional GAN architecture together with the successful application to inverse design of materials would inspire further exploration in research fields beyond materials. (C) 2020 Elsevier Ltd. All rights reserved.