Perspective
Automated analysis of pore structures in biomaterials
Schematic representation of factors affecting automated pore size determination and requirements for improvement.
Bioinspired rational design of nanozymes
This review presents a comprehensive overview of bioinspired rational design of nanozymes, guided by the catalytic mechanisms and structural characteristics of natural enzymes, and enhanced by emerging trend of machine learning assisted strategies.
Establishing baselines for generative discovery of inorganic crystals
This paper compares generative AI models with template-based methods for discovering new materials, evaluating their ability to generate stable structures and target desired properties.
Communication
Heat transport properties of PbTe1−xSex alloys using equivariant graph neural network interatomic potential
The suppression of heat transport in disordered crystals arises from a competition between mass fluctuations and bond disorder, but their relative contributions remain difficult to disentangle.
Communication
Deep learning-enhanced development of innovative antioxidant liposomal drug delivery systems from natural herbs
Discovery and construction of antioxidant-liposomal platforms from natural herbs based on the BERT model and fully validated in vitro and in vivo.
Communication
Experiment-in-loop interactive optimization of polymer composites for “5G-and-beyond” communication technologies
The fabrication of high-performance polymer composites for “5G-and-beyond” communication technologies was achieved through experiment-in-loop optimization facilitated by ARD kernel-equipped Bayesian optimization.
Communication
Machine-learning accelerated prediction of two-dimensional conventional superconductors
We perform a large scale search for two-dimensional (2D) superconductors, by using electron–phonon calculations with density-functional perturbation theory combined with machine learning models.
Paper
Predicting neutron experiments from first principles: a workflow powered by machine learning
From electronic structure calculations via molecular dynamics to neutron spectra.
On-demand design of materials with enhanced dielectric properties via a machine learning-assisted materials genome approach
An on-demand design of silicon-containing arylacetylene resins with specific thermal and dielectric performance was achieved via the machine learning-assisted materials genome approach.
Paper
Can large language models predict the hydrophobicity of metal–organic frameworks?
Fine-tuning a large language model to predict the hydrophobicity of metal–organic frameworks.
Paper
Sparse modeling based Bayesian optimization for experimental design
Efficiently optimizing high-dimensional synthesis parameters via Bayesian optimization (BO) is a challenge in recent materials exploration. This study introduces a sparse modeling-based BO method using the maximum partial dependence effect (MPDE).
Paper
Symmetry-informed graph neural networks for carbon dioxide isotherm and adsorption prediction in aluminum-substituted zeolites
SymGNN leverages crystal symmetries to improve adsorption predictions in zeolites. By encoding symmetry-aware features, the model achieves accurate isotherm and heat of adsorption predictions, even for unseen topologies.
Data-driven design and green preparation of bio-based flame retardant polyamide composites
This work presents a bio-based flame retardant for biocomposites, using data-driven experimentation and optimization to boost tensile strength by 22.3% and reduce heat release rate by 73.7% vs. the neat polymer.
About this collection
Automation, machine learning, and artificial intelligence has enabled new frontiers in materials discovery. The advent of high-throughput calculations and the creation of comprehensive materials property databases, machine learning and AI are now equipped to navigate vast compositional spaces and effortlessly and rapidly predict physical properties. At the same time, automated laboratories are driving a revolution in synthetic methodologies. However, amidst these transformative developments lie specific challenges intrinsic to the autonomous discovery of materials.
This cross-journal collection published in Materials Horizons, Journal of Materials Chemistry A, B and C and Materials Advances explores this theme of Frontiers in materials discovery – innovations and challenges in machine learning and artificial intelligence.
Guest Edited by DrJakoah Brgoch (University of Houston, USA), Dr Alex Ganose (Imperial College London, UK), Professor Janine George (Federal Institute of Materials Research and Testing (BAM), Germany, and University of Jena, Germany), Dr Kedar Hippalgaonkar (Nanyang Technological University, Singapore), Professor David Scanlon (University of Birmingham, UK), the collection convenes innovative research spanning various disciplines, including materials science, chemistry, physics, engineering, computer science, statistics, and robotics, with the aim of stimulating discussions and fostering novel collaborations. These interdisciplinary connections are essential for developing new AI-driven algorithms, automated processes, and human-robot collaborations crucial to enhancing the data-driven scientific workflow necessary for material discovery.
We hope you enjoy reading the papers featured in the collection.