Machine learning applied to the design and optimization of polymeric materials: A review

Polymeric materials have long been at the forefront of materials science due to their exceptional physical and chemical properties. Designing new polymeric materials traditionally relies on a combination of physical and chemical principles along with empirical trial and error methods. A big issue in this approach is that the traditional trial and error experiments/simulations are often time consuming and sometimes inefficient in order to achieve targeted properties. Machine learning (ML) offers a promising solution to this issue. By leveraging ML algorithms, researchers can accelerate the design process, predict material properties more rapidly, and optimize formulations. ML approaches can analyze vast amounts of data, uncover hidden patterns, and generate predictive models that significantly reduce the time needed to develop materials with desired properties. This paradigm shift allows for more efficient exploration of the material space and faster innovation in polymeric material design. The current review article focuses on the application of ML methods to help design polymeric materials, more especifically, thermosets, thermoplastics and elastomers. Additionally, the review explores future possibilities and opportunities arising from the state-of-the-art advancements in ML, offering perspectives on how these technologies might evolve and influence the polymer science domain. • Tailored ML strategies for thermosets, thermoplastics, and elastomers are provided in this review. • The key role of ML in optimizing formulations and accelerating polymer design is discussed. • Opportunities and challenges are highlighted, encouraging advancements in ML for polymer science.