Article
|
Open Access
Featured
-
-
Article
| Open AccessA magnetically actuated dynamic labyrinthine transmissive ultrasonic metamaterial
Space-coiling acoustic metamaterials, where sound travels through labyrinthine geometries, are interesting for their high energy transmission and broad modulation characteristics. Here, the authors demonstrate an active approach to acoustic metamaterial reconfiguration based on dynamic space-coiling unit cells and soft robotic-inspired actuators.
- Christabel Choi
- , Shubhi Bansal
- & Sriram Subramanian
-
Article
| Open AccessMachine learning-derived reaction statistics for 3D spectroimaging of copper sulfidation in heterogeneous rubber/brass composites
Copper sulfidation in the rubber/brass interface of tires during aging is still not well understood. Here, the 3D spatial location and chemical states of copper species in a rubber/brass composite are visualized and tracked by 3D X-ray spectroimaging with data-driven machine learning analysis.
- Hirosuke Matsui
- , Yuta Muramoto
- & Mizuki Tada
-
Article
| Open AccessA multi-timescale synaptic weight based on ferroelectric hafnium zirconium oxide
Brain-inspired neuromorphic computing is a key technology for processing an ever-growing amount of data. Here, an artificial synapse with dual resistance modulation mechanisms is demonstrated, achieving a dynamic range of 60, an endurance exceeding 1010 cycles, and more than 10 years of retention.
- Mattia Halter
- , Laura Bégon-Lours
- & Bert Jan Offrein
-
Review Article
| Open AccessGraph neural networks for materials science and chemistry
Graph neural networks are machine learning models that directly access the structural representation of molecules and materials. This Review discusses state-of-the-art architectures and applications of graph neural networks in materials science and chemistry, indicating a possible road-map for their further development.
- Patrick Reiser
- , Marlen Neubert
- & Pascal Friederich
-
Article
| Open AccessIntegration of machine learning with neutron scattering for the Hamiltonian tuning of spin ice under pressure
Designing and understanding quantum materials requires continuous feedback between experimental observations and theoretical modelling. Here, a machine learning scheme integrates experiments with theory and modelling on experimental timescales for extracting material parameters and properties of Dy2Ti2O7 spin-ice under pressure.
- Anjana Samarakoon
- , D. Alan Tennant
- & Santiago A. Grigera
-
Perspective
| Open AccessWhy big data and compute are not necessarily the path to big materials science
Machine learning is an increasingly important tool for materials science. Here, the authors suggest that its contextual use, including careful assessment of resources and bias, judicious model selection, and an understanding of its limitations, will help researchers to expedite scientific discovery.
- Naohiro Fujinuma
- , Brian DeCost
- & Samuel E. Lofland
-
Article
| Open AccessAccelerating error correction in tomographic reconstruction
Recent advances in scanning probe-based tomographic imaging have greatly improved spatial resolution, but systematic and random errors are a serious impediment to reliable data extraction. Here, a combined optimization and alignment algorithm provides a scalable approach to error-correcting reconstruction of large datasets.
- Sajid Ali
- , Matthew Otten
- & Z. W. Di
-
Article
| Open AccessDeep learning for the rare-event rational design of 3D printed multi-material mechanical metamaterials
Multi-material 3D printing techniques are now enabling the rational design of metamaterials with both complex geometries and multiple materials compositions. Here, deep-learning methods are used to identify, among planar network structures, the rare designs that yield very unusual and desirable combinations of materials properties.
- Helda Pahlavani
- , Muhamad Amani
- & Amir A. Zadpoor
-
Article
| Open AccessAlgorithmic design of origami mechanisms and tessellations
Origami is a promising source of inspiration in designing foldable structures and reconfigurable metamaterials. Here, building on exact folding kinematic conditions, an algorithmic design of rigidly-foldable origami structures is presented, allowing the engineering of metamaterials with arbitrary complex crease patterns.
- Andreas Walker
- & Tino Stankovic
-
Article
| Open AccessTracking the time evolution of soft matter systems via topological structural heterogeneity
Topological data analysis is an important framework for quantifying the structural and morphological features of soft materials. Here, structural heterogeneity is introduced as a quantitative measure of non-equilibrium mesoscopic order in soft matter and used to track the time-evolution of liquid crystal phase transitions.
- Ingrid Membrillo Solis
- , Tetiana Orlova
- & Malgosia Kaczmarek
-
Article
| Open AccessPredicting synthesizability of crystalline materials via deep learning
Predicting the synthesizability of unknown crystals is important for accelerating materials discovery. Here, the synthesizability of crystals with any given composition and structure can be predicted by a deep learning model that maps crystals onto color-coded 3D images processed by convolutional neural networks.
- Ali Davariashtiyani
- , Zahra Kadkhodaie
- & Sara Kadkhodaei
-
Article
| Open AccessA geometric-information-enhanced crystal graph network for predicting properties of materials
Graph neural networks are an accurate machine learning-based approach for property prediction. Here, a geometric-information-enhanced crystal graph neural network is demonstrated, which accurately predicts the formation energy and band gap of crystalline materials.
- Jiucheng Cheng
- , Chunkai Zhang
- & Lifeng Dong
-
Article
| Open AccessMachine learning autonomous identification of magnetic alloys beyond the Slater-Pauling limit
Finding materials with large magnetization is highly desirable for technological applications. Here, a machine learning autonomous search and automated combinatorial synthesis reveal that multi-element alloys with Ir and Pt impurities have a magnetization exceeding the Slater-Pauling limit of Fe75Co25.
- Yuma Iwasaki
- , Ryohto Sawada
- & Masahiko Ishida
-
Article
| Open AccessIntegrating multiple materials science projects in a single neural network
Traditionally, machine learning for materials science is based on database-specific models and is limited in the number of predictable parameters. Here, a versatile graph-based neural network can integrate multiple data sources, allowing the prediction of more than 40 parameters simultaneously.
- Kan Hatakeyama-Sato
- & Kenichi Oyaizu
-
Article
| Open AccessThe +2 oxidation state of Cr incorporated into the crystal lattice of UO2
Uranium dioxide is commonly doped with chromium to improve its performance as a nuclear fuel. Here, with the aid of ab initio simulations and re-evaluation of experimental data, the oxidation state of chromium in the uranium dioxide lattice is identified as +2, not the widely believed +3.
- Mengli Sun
- , Joshua Stackhouse
- & Piotr M. Kowalski