arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2605.30195cond-mat.mtrl-scics.AIcs.LG
收藏

分子MPNN性能驱动因素:算子级因子基准测试

What drives performance in molecular MPNNs? An operator-level factorial benchmark

Materials Genome Institute, Shanghai University(上海大学材料基因组研究所);School of Computer Engineering and Science, Shanghai University(上海大学计算机工程与科学学院);School of Materials Science and Engineering, Tongji University(同济大学材料科学与工程学院)

Panyu Jiao, Shuizhou Chen, Yiheng Shen, Yuyang Wang, Runhai Ouyang, Wei Xie

AI总结:

通过分解分子MPNN为消息种子初始化、节点-边融合和节点更新三类算子,在84种配置下对MoleculeNet数据集进行基准测试,发现消息构建而非更新复杂度主导性能,并提出了设计启发式方法。

摘要:

消息传递神经网络(MPNN)广泛用于分子性质预测,但其作为整体架构部署使得难以识别特定消息传递算子如何影响性能。我们提出了一个算子级因子基准测试,将二维分子MPNN分解为消息种子初始化、节点-边融合和节点更新算子三个家族。在共享实验设置和统计分析协议下,对十个MoleculeNet数据集上的84种配置进行了基准测试。在这个受控设计中,性能变化主要与消息构建相关,而非更新复杂度。消息种子初始化在回归和分类任务中均显示出显著的家族级效应;节点-边融合在回归任务中显示出显著的家族级效应,且基于拼接的混合具有描述性优势;更新家族在任一任务家族中均未显示出统计上支持的效应。对Quinethazone分子的表征探测进一步表明,与Hadamard门控相比,基于拼接的混合能更好地区分化学上不同的杂原子并抵抗过度平滑。分别针对分类和回归任务选择的代表性配置相对于已建立的分子图神经网络(GNN)基线恢复了竞争性性能,在十个基准数据集中有八个数值上排名最佳。这些实证结果通过对代表性节点-边融合和更新算子的简洁机理分析进行了解释。我们的发现通过将模型设计从搜索整体架构转变为针对化学信息在消息传递管道中进入位置和方式的定向评估,为分子MPNN提供了实证设计启发式方法。

英文摘要:

Message-passing neural networks (MPNNs) are widely used for molecular property prediction, but their deployment as monolithic architectures makes it difficult to identify how specific message-passing operators affect performance. We present an operator-level factorial benchmark that decomposes 2D molecular MPNNs into the three families of message-seed initialization, node-edge fusion, and node update operators. The resulting 84 configurations are benchmarked on ten MoleculeNet datasets under a shared experimental setup and statistical analysis protocol. Across this controlled design, performance variation is associated primarily with message construction rather than update complexity. Message-seed initialization shows significant family-level effects for both regression and classification, node-edge fusion shows a significant family-level effect for regression with descriptive advantages for concatenation-based mixing, and the update family shows no statistically supported effect for either endpoint family. A representation probe into the Quinethazone molecule further demonstrates that concatenation-based mixing can better differentiate chemically distinct heteroatoms and withstand oversmoothing than Hadamard gating. Representative configurations selected separately for classification and regression recover competitive performance relative to established molecular graph neural network (GNN) baselines, ranking numerically best on eight of ten benchmark datasets. These empirical results are interpreted through concise mechanistic analyses of representative node-edge fusion and update operators. Our findings provide empirical design heuristics for molecular MPNNs by turning model design from a search over monolithic architectures into a targeted assessment of where and how chemical information enters the message-passing pipeline.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑