Search for: All records

Abstract High-throughput materials research is strongly required to accelerate the development of safe and high energy-density lithium-ion battery (LIB) applicable to electric vehicle and energy storage system. The artificial intelligence, including machine learning with neural networks such as Boltzmann neural networks and convolutional neural networks (CNN), is a powerful tool to explore next-generation electrode materials and functional additives. In this paper, we develop a prediction model that classifies the major composition (e.g., 333, 523, 622, and 811) and different states (e.g., pristine, pre-cycled, and 100 times cycled) of various Li(Ni, Co, Mn)O₂(NCM) cathodes via CNN trained on scanning electron microscopy (SEM) images. Based on those results, our trained CNN model shows a high accuracy of 99.6% where the number of test set is 3840. In addition, the model can be applied to the case of untrained SEM data of NCM cathodes with functional electrolyte additives. 

Abstract Direct electron detectors in scanning transmission electron microscopy give unprecedented possibilities for structure analysis at the nanoscale. In electronic and quantum materials, this new capability gives access to, for example, emergent chiral structures and symmetry-breaking distortions that underpin functional properties. Quantifying nanoscale structural features with statistical significance, however, is complicated by the subtleties of dynamic diffraction and coexisting contrast mechanisms, which often results in a low signal-to-noise ratio and the superposition of multiple signals that are challenging to deconvolute. Here we apply scanning electron diffraction to explore local polar distortions in the uniaxial ferroelectric Er(Mn,Ti)O₃. Using a custom-designed convolutional autoencoder with bespoke regularization, we demonstrate that subtle variations in the scattering signatures of ferroelectric domains, domain walls, and vortex textures can readily be disentangled with statistical significance and separated from extrinsic contributions due to, e.g., variations in specimen thickness or bending. The work demonstrates a pathway to quantitatively measure symmetry-breaking distortions across large areas, mapping structural changes at interfaces and topological structures with nanoscale spatial resolution. 

Abstract Ferroelectrics, due to their polar nature and reversible switching, can be used to dynamically control surface chemistry for catalysis, chemical switching, and other applications such as water splitting. However, this is a complex phenomenon where ferroelectric domain orientation and switching are intimately linked to surface charges. In this work, the temperature‐induced domain behavior of ferroelectric‐ferroelastic domains in free‐standing BaTiO₃films under different gas environments, including vacuum and oxygen‐rich, is studied by in situ scanning transmission electron microscopy (STEM). An automated pathway to statistically disentangle and detect domain structure transformations using deep autoencoders, providing a pathway towards real‐time analysis is also established. These results show a clear difference in the temperature at which phase transition occurs and the domain behavior between various environments, with a peculiar domain reconfiguration at low temperatures, from a‐c to a‐a at ≈60 °C. The vacuum environment exhibits a rich domain structure, while under the oxidizing environment, the domain structure is largely suppressed. The direct visualization provided by in situ gas and heating STEM allows to investigate the influence of external variables such as gas, pressure, and temperature, on oxide surfaces in a dynamic manner, providing invaluable insights into the intricate surface‐screening mechanisms in ferroelectrics. 

Label-free cell classification is advantageous for supplying pristine cells for further use or examination, yet existing techniques frequently fall short in terms of specificity and speed. In this study, we address these limitations through the development of a novel machine learning framework, Multiplex Image Machine Learning (MIML). This architecture uniquely combines label-free cell images with biomechanical property data, harnessing the vast, often underutilized biophysical information intrinsic to each cell. By integrating both types of data, our model offers a holistic understanding of cellular properties, utilizing cell biomechanical information typically discarded in traditional machine learning models. This approach has led to a remarkable 98.3% accuracy in cell classification, a substantial improvement over models that rely solely on image data. MIML has been proven effective in classifying white blood cells and tumor cells, with potential for broader application due to its inherent flexibility and transfer learning capability. It is particularly effective for cells with similar morphology but distinct biomechanical properties. This innovative approach has significant implications across various fields, from advancing disease diagnostics to understanding cellular behavior. 

Retinal image plays a crucial role in diagnosing various diseases, as retinal structures provide essential diagnostic information. However, effectively capturing structural features while integrating them with contextual information from retinal images remains a challenge. In this work, we propose segmentation-guided dual-branch network for retinal disease diagnosis using retinal images and their segmentation maps, named SegImgNet. SegImgNet incorporates a segmentation module to generate multi-scale retinal structural feature maps from retinal images. The classification module employs two encoders to independently extract features from segmented images and retinal images for disease classification. To further enhance feature extraction, we introduce the Segmentation-Guided Attention (SGA) block, which leverages feature maps from the segmentation module to refine the classification process. We evaluate SegImgNet on the public AIROGS dataset and the private e-ROP dataset. Experimental results demonstrate that SegImgNet consistently outperforms existing methods, underscoring its effectiveness in retinal disease diagnosis.