Machine Learning Approaches in Polymer Science: Progress and Fundamental for a New Paradigm

SmartMat, Год журнала: 2025, Номер 6(1)

Опубликована: Янв. 9, 2025

ABSTRACT Machine learning (ML), material genome, and big data approaches are highly overlapped in their strategies, algorithms, models. They can target various definitions, distributions, correlations of concerned physical parameters given polymer systems, have expanding applications as a new paradigm indispensable to conventional ones. Their inherent advantages building quantitative multivariate largely enhanced the capability scientific understanding discoveries, thus facilitating mechanism exploration, prediction, high‐throughput screening, optimization, rational inverse designs. This article summarizes representative progress recent two decades focusing on design, preparation, application, sustainable development materials based exploration key composition–process–structure–property–performance relationship. The integration both data‐driven insights through ML deepen fundamental discover novel is categorically presented. Despite construction application robust models, strategies algorithms deal with variant tasks science still rapid growth. challenges prospects then We believe that innovation will thrive along approaches, from efficient design applications.

Язык: Английский

---

Machine Learning in Polymer Research

Advanced Materials, Год журнала: 2025, Номер unknown

Опубликована: Фев. 9, 2025

Machine learning is increasingly being applied in polymer chemistry to link chemical structures macroscopic properties of polymers and identify patterns the that help improve specific properties. To facilitate this, a dataset needs be translated into machine readable descriptors. However, limited inadequately curated datasets, broad molecular weight distributions, irregular configurations pose significant challenges. Most off shelf mathematical models often need refinement for applications. Addressing these challenges demand close collaboration between chemists mathematicians as must formulate research questions terms while are required refine This review unites both disciplines address curation hurdles highlight advances synthesis modeling enhance data availability. It then surveys ML approaches used predict solid-state properties, solution behavior, composite performance, emerging applications such drug delivery polymer-biology interface. A perspective field concluded importance FAIR (findability, accessibility, interoperability, reusability) integration theory discussed, thoughts on machine-human interface shared.

Язык: Английский

---

On-demand reverse design of polymers with PolyTAO

npj Computational Materials, Год журнала: 2024, Номер 10(1)

Опубликована: Ноя. 29, 2024

The forward screening and reverse design of drug molecules, inorganic polymers with enhanced properties are vital for accelerating the transition from laboratory research to market application. Specifically, due scarcity large-scale datasets, discovery via materials informatics is particularly challenging. Nonetheless, scientists have developed various machine learning models polymer structure-property relationships using only small thereby advancing process polymers. However, success this approach ultimately depends on diversity candidate pool, exhaustively enumerating all possible structures through human imagination impractical. Consequently, achieving on-demand essential. In work, we curate an immense dataset containing nearly one million polymeric pairs based expert knowledge. Leveraging dataset, propose a Transformer-Assisted Oriented pretrained model generation (PolyTAO). This generates 99.27% chemical validity in top-1 mode (approximately 200k generated polymers), representing highest reported rate among generative models, was achieved largest test set. Importantly, average R2 between their expected values across 15 predefined 0.96, which underscores PolyTAO's powerful capabilities. To further evaluate model's performance generating additional user-defined downstream tasks, conduct fine-tuning experiments three publicly available datasets both semi-template template-free paradigms. Through these extensive experiments, demonstrate that our its fine-tuned versions capable specified properties, whether or more challenging scenarios, showcasing potential as unified foundation generation.

Язык: Английский

---

Machine Learning‐Assisted Design of Multilayer Thermoplastic Composites: Robust Neural Network Prediction and Feature Importance Analysis

Macromolecular Materials and Engineering, Год журнала: 2025, Номер unknown

Опубликована: Май 13, 2025

Abstract Multilayer thermoplastic composites offer sustainable alternatives to traditional thermoset and metal materials. However, their design is inherently complex, involving numerous interdependent parameters that render conventional processes both expensive time‐consuming. While machine learning‐assisted methods provide a potential solution, they typically require large datasets can be costly obtain. This study explores robust neural network, specifically, an Advanced Perceptron (AdvMLP) Regressor, predict the peel strength of multilayer composites. Through architectural enhancements, AdvMLP effectively trained on limited yet authentic manufacturing dataset, yielding predictions validated by benchmark metrics k‐fold cross‐validation. The model captures intricate interplay between composite properties, enabling comprehensive feature importance analysis dimensionality reduction. Overall, this establishes generalizable methodology guide accelerate optimization

Язык: Английский

---

Data Science-Centric Design, Discovery, and Evaluation of Novel Synthetically Accessible Polyimides with Desired Dielectric Constant

Chemical Science, Год журнала: 2024, Номер unknown

Опубликована: Янв. 1, 2024

A data-science-centered “design–discover–evaluate” scheme is presented, and 9 novel polyimides suitable for application to high-temperature energy storage dielectrics are identified from the designed virtual structure library.

Язык: Английский

---

Application of machine learning in polyimide structure design and property regulation

High Performance Polymers, Год журнала: 2025, Номер unknown

Опубликована: Янв. 8, 2025

Polyimide (PI) is widely used in modern industry due to its excellent properties. Its synthesis methods and property research have significantly progressed. However, the design regulation of PI structures through traditional technologies are slow expensive, which make it difficult meet practical demand materials. With rapid development high-throughput computing data-driven technology, machine learning (ML) has become an important method for exploring new Data-driven ML envisaged as a decisive enabler PIs discovery. This paper first introduces basic workflow common algorithms ML. Secondly, applications material properties prediction, assisting computational simulation inverse desired reviewed. Finally, we discuss main challenges possible solutions research.

Язык: Английский

---

Exceptionally High-Temperature-Resistant Kapton-Type Polyimides with Tg > 520 °C: Synthesis via Incorporation of Spirobis(indene)-bis(benzoxazole)-Containing Diamines

Polymers, Год журнала: 2025, Номер 17(7), С. 832 - 832

Опубликована: Март 21, 2025

Polyimides (PIs), recognized for their exceptional thermal stability, are extensively employed in advanced applications, including aerospace, flexible displays, solar cells, flame-retardant materials, and high-temperature filtration materials. However, with the continuous advancements science technology, demand improved performance of PIs these application areas has increased significantly. In this study, four spirobis(indene)-bis(benzoxazole) diamine monomers (5a, 5aa, 5b 5bb) were designed synthesized. These copolymerized pyromellitic dianhydride (PMDA) 4,4-diaminodiphenylmethane (ODA) to develop Kapton-type PIs. By varying copolymerization molar ratios different diamines, a series novel ultrahigh-temperature-resistant PI films successfully prepared, it was found that incorporating highly rigid twisted structure into matrix enhances rigidity polymer chains restricts mobility, thereby significantly improving films. When 5a ODA at 1:9 4:6, glass transition temperature (Tg) from 396 °C 467 >520 °C, respectively. also exhibit mechanical properties, modulus increasing 1.6 GPa 4.7 GPa, while demonstrating low dielectric performance, as evidenced by decrease constant (Dk) 3.51 3.08 under 10 GHz high-frequency electric field. Additionally, molecular dynamics simulations further explore relationships between structure, condensed states, film providing theoretical guidance development polymers ultrahigh resistance superior overall performance.

Язык: Английский

---

Algorithm in chemistry: molecular logic gate-based data protection

Chemical Society Reviews, Год журнала: 2025, Номер unknown

Опубликована: Янв. 1, 2025

This review focuses on stimuli-responsive material (SRM)-based data protection, emphasizing the integration of intricate logic and algorithms in SRM-constructed hardware. It also discusses current challenges future directions field.

Язык: Английский

---

Graph Neural Networks for Surface Tension Prediction of Polymers: A Comparative Analysis with Descriptor-Based Models

Macromolecules, Год журнала: 2025, Номер unknown

Опубликована: Апрель 13, 2025

Язык: Английский

---

Enclose Biobased Content into Polyurethane Elastomers: A Summary of Synthetic Routes and an Inverse Prediction of their Percentages

Macromolecular Rapid Communications, Год журнала: 2025, Номер unknown

Опубликована: Апрель 28, 2025

Abstract Biobased polyurethane elastomers (BPUEs) got vigorous exploration in recent decade, clarifying how biobased content affect their mechanical and thermal properties becomes critical. Here, a comprehensive BPUE dataset with 506 splines associated brief summary for the transformation of into are presented. Distributions typical including Young's modulus, tensile strength, elongation at break, glass transition temperature (T g ) clarified. A linear relationship T DMA = 0.98*T + 19.43K is found ’s measured by differential scanning calorimetry (DSC) dynamic analysis (DMA). Then, binary classification model an accuracy 0.80 to distinguish PUEs without content, regression coefficient determination 0.89 predict mass percentage (BBC%) constructed. Based on these predictive models, important correlations observed, strong dependence Tg tanδ BBC%, weak preference use longer spline test, lower heating rates DSC measurements BPUEs higher BBC%. These inverse models provide cost‐effective way quantify PUE products, prior knowledge exact composition formulas.

Язык: Английский

---