[论文] ZeoBench: A Benchmark for Self-Supervised Learning on 3D Zeolite Repre...

论文概要

研究领域: CV
作者: Aaron Sun, Yachan Liu, Ping Yang, Peng Bai, Subhransu Maji
发布时间: 2026-11-06
arXiv: 2611.09880

中文摘要

沸石是一类重要的多孔晶体材料，在气体储存、分离和催化中应用广泛。尽管已知数十万种实验实现和假设的沸石结构，但为目标应用识别优化材料仍然困难——属性标签获取昂贵且仅对有限的结构-分子对可用。先前工作已证明3D神经表示能有效预测沸石属性，但在材料发现相关的低数据情形下，哪些表示学习策略最有效仍不清楚。本文提出 ZeoBench——面向全硅沸石稠密3D表示的自监督学习基准，并在六个现有下游吸附性质预测任务和一系列训练集规模上评估所学特征。研究比较了沸石体积数据的自监督3D预训练、通过多视图和通道膨胀策略将图像预训练的2D ConvNet和ViT适配到3D、在域外体积数据集上预训练的3D ConvNet，以及手工描述符和监督基线。结果表明，自监督3D表示优于现有方法，在域内标签机制和未见的金属有机框架域上都达到最先进水平。学习表示一致优于手工描述符，卷积架构在此场景下通常比视觉Transformer更有效。

原文摘要

Zeolites are an important class of porous crystalline materials with broad applications in gas storage, separations, and catalysis. Although hundreds of thousands of experimentally realized and hypothetical zeolite structures are known, identifying materials optimized for a target application remains challenging because property labels are expensive to obtain and available only for limited structure-molecule pairs. While prior work has demonstrated that 3D neural representations can predict zeolite properties effectively, it remains unclear which representation-learning strategies are most effective in the low-data regime relevant to materials discovery. We present ZeoBench, a benchmark for self-supervised learning on dense 3D representations of all-silica zeolites, and evaluate the result...

自动采集于 2026-11-10

#论文 #arXiv #CV #小凯