用等变图神经网络提升材料光学谱预测精度，助力光电材料筛选。
Equivariant Graph Neural Networks Improve Optical Spectra Prediction for Materials Screening
- 采用等变图神经网络建模材料结构，保留几何信息
- 在10,533个结构上实现优于现有方法的预测性能
- 尤其擅长0-8电子伏特范围和静态介电常数预测
高通量材料筛选中光学谱的可扩展预测对光电应用（如太阳能电池）至关重要。现有代理模型通常基于较低理论层次计算的光谱训练，或依赖旋转不变的标量特征，限制了其几何表达能力。本文探索将等变图神经网络用于光学谱预测，改进GotenNet模型，并在多个数据集上评估，包括最新发布的10,533个结构、光谱基于随机相位近似（RPA）计算的数据集。所提模型显著超越当前最先进水平，尤其在0-8 eV能量范围内及静态实介电常数预测上表现最佳，这对薄膜光学具有重要意义。
原文摘要 · Abstract (English)
Scalable prediction of optical spectra is a critical component of high-throughput materials screening for optoelectronic applications such as solar cells. Existing surrogate models are trained on spectra computed from lower levels of theory or rely on rotation-invariant scalar features, limiting their geometric expressiveness. We explore the use of equivariant graph neural networks for optical spectra prediction, adapting GotenNet to this task and evaluating it on multiple datasets including a recently published collection of 10,533 structures with spectra computed at the level of the random phase approximation (RPA). The proposed model outperforms the current state of the art, with the largest gains in the 0-8 eV range and on predicting the static real permittivity, both of particular relevance for thin-film optics.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。