This is the authors' abstract. We don't add key points for this paper.
Olefin polymerization is a cornerstone of the modern chemical industry, and the development of high-performance metal catalysts is essential for controlling polymer microstructure and properties. Traditionally, catalyst optimization has relied on resource-intensive experimental screening. However, the past decade has seen the emergence of machine learning as a transformative tool for identifying non-linear patterns in highdimensional datasets and accelerating catalyst discovery. This review discusses recent efforts in applying supervised, unsupervised, and active learning methods to metal-catalyzed olefin polymerization. We examine the development of molecular descriptors—both manually crafted from density functional theory and learned through graph neural networks—and outline typical workflows for data curation, model selection, and interpretability using tools like Shapley additive explanations. Key applications are highlighted across ethylene and propylene homopolymerization, oligomerization, and copolymerization with polar monomers, demonstrating how machine learning can predict catalytic activity, molecular weight, and selectivity while providing mechanistic insights. Finally, we discuss the current challenges and future opportunities for data-driven catalysis research.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Wong et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: