[论文解读] Intelligent Materials Modelling: Large Language Models Versus Partial Least Squares Regression for Predicting Polysulfone Membrane Mechanical Performance

Dingding Cao, Mieow Kee Chan+3 more|arXiv (Cornell University)|Mar 14, 2026

Machine Learning in Materials Science被引用 0

一句话总结

该研究将四种大语言模型（LLM）与偏最小二乘回归（PLS）在从结构描述符预测聚砜膜（PSF）力学性能上的任务进行基准评估，发现LLM尤以提高断裂伸长率（EL）为显著，而对杨氏模量（E）和拉伸强度（TS）等线性性质，PLS仍具竞争力。

ABSTRACT

Predicting the mechanical properties of polysulfone (PSF) membranes from structural descriptors remains challenging due to extreme data scarcity typical of experimental studies. To investigate this issue, this study benchmarked knowledge-driven inference using four large language models (LLMs) (DeepSeek-V3, DeepSeek-R1, ChatGPT-4o, and GPT-5) against partial least squares (PLS) regression for predicting Young's modulus (E), tensile strength (TS), and elongation at break (EL) based on pore diameter (PD), contact angle (CA), thickness (T), and porosity (P) measurements. These knowledge-driven approaches demonstrated property-specific advantages over the chemometric baseline. For EL, LLMs achieved statistically significant improvements, with DeepSeek-R1 and GPT-5 delivering 40.5% and 40.3% of Root Mean Square Error reductions, respectively, reducing mean absolute errors from $11.63\pm5.34$% to $5.18\pm0.17$%. Run-to-run variability was markedly compressed for LLMs ($\leq$3%) compared to PLS (up to 47%). E and TS predictions showed statistical parity between approaches ($q\geq0.05$), indicating sufficient performance of linear methods for properties with strong structure-property correlations. Error topology analysis revealed systematic regression-to-the-mean behavior dominated by data-regime effects rather than model-family limitations. These findings establish that LLMs excel for non-linear, constraint-sensitive properties under bootstrap instability, while PLS remains competitive for linear relationships requiring interpretable latent-variable decompositions. The demonstrated complementarity suggests hybrid architectures leveraging LLM-encoded knowledge within interpretable frameworks may optimise small-data materials discovery.

研究动机与目标

• 研究从稀疏实验数据中预测PSF膜力学性能的挑战。
• 比较基于PD、CA、T、P的四种LLM的知识驱动推断与PLS回归在E、TS、EL方面的表现。
• 评估运行-到-运行的变异性和误差特征，以理解自助法不稳定性下的模型可靠性。
• 识别特定性质的优势以及材料发现中的混合、可解释架构潜力。

提出的方法

• 评估四种LLM（DeepSeek-V3、DeepSeek-R1、ChatGPT-4o、GPT-5）相对于PLS回归。
• 从孔径PD、接触角CA、厚度T和孔隙度P预测杨氏模量E、拉伸强度TS和断裂伸长EL。
• 计算性能指标（如RMSE、MAE）并比较统计显著性（如通过q值）。
• 分析误差拓扑以理解回归均值化效应和数据-体制的影响。

实验结果

研究问题

• RQ1 LLMs在从有限数据预测PSF膜力学性能方面是否优于PLS？
• RQ2 哪些性质（E、TS、EL）在LLMs相对于PLS方面得到最显著的提升？
• RQ3 在自助法不稳定性下，LLMs与PLS的运行-到-运行变异性如何比较？
• RQ4 误差拓扑揭示的数据-体制效应与模型族限制之间的关系？
• RQ5 是否可以通过混合的可解释LLM驱动框架进一步提升小数据材料发现？

主要发现

• LLMs在特定性质方面具有优势，尤其是EL方面，DeepSeek-R1（40.5%）与GPT-5（40.3%）在RMSE上显著降低。
• EL的平均绝对误差从11.63±5.34%降至5.18±0.17%（LLMs）。
• LLMs显示明显更低的运行-到-运行变异性（≤3%），相比之下PLS最高可达47%。
• E与TS的预测在LLMs与PLS之间统计上相同（q≥0.05）。
• 误差拓扑指示回归到均值的行为由数据-体制效应驱动而非固有模型局限；LLMs在非线性、对约束敏感的性质上表现出色，而PLS在线性关系上仍具竞争力。