Search

Deep learning has catalyzed a transformative shift in material discovery, offering a key advantage over traditional experimental and theoretical methods by significantly reducing associated costs. Models adept at predicting properties from chemical compositions alone do not require structural information. However, this cost-efficient approach compromises model precision, particularly in Chemical Composition-based Property Prediction Models (CPMs), which are notably less accurate than Structure-based Property Prediction Models (SPMs). Addressing this challenge, our study introduces a novel Teacher-Student (TS) strategy, where a pretrained SPM serves as an instructive ‘teacher’ to enhance the CPM’s precision. This TS strategy successfully harmonizes low-cost exploration with high accuracy, achieving a significant 47.1% reduction in relative error in scenarios involving 100 data entries. We also evaluate the effectiveness of the proposed strategy by employing perovskites as a case study. This method represents a significant advancement in the exploration and identification of valuable materials, leveraging CPM’s potential while overcoming its precision limitations.

Liquid crystal elastomers (LCEs) are a kind of polymer network that combines the entropic elasticity of polymer networks and the mesogenic unit by means of mild cross-linking. LCEs have been extensively investigated in various fields, including artificial muscles, actuators, and microrobots. However, LCEs are characterized by the poor mechanical properties of the light polymers themselves. In this study, we propose to prepare a carbon nanotube/liquid crystal elastomer (CNT/LCE) composite yarn by electrospinning technology and a two-step cross-linking strategy. The CNT/LCE composite yarn exhibits a reversible shrinkage ratio of nearly 70%, a tensile strength of 16.45 MPa, and a relatively sensitive response speed of ∼3 s, enabling a fast response by photothermal actuation. The research disclosed in this article may provide new insights for the development of artificial muscles and next-generation smart robots.

In a high-dimensional sample space, the distribution characteristics of samples are complex. It is difficult to accurately describe the local distribution characteristics of samples directly in the original sample space. This paper proposes a multiple discriminant preserving support subspace radial basis function neural networks with graph similarity learning (DPSS-RBFNN) to describe the distribution characteristics of samples in the original high-dimensional space by multiple low-dimensional and simple discriminant preserving subspaces. DPSS-RBFNN includes the discriminant preserving support subspace (DPSS) learning module and the subspace distribution feature extraction (SDFE) module. In the DPSS learning module, the discriminativeness of each attribute and the joint discriminant between attributes are first considered to construct multiple subspaces. The discriminativeness of the features in these subspaces is not lower than that of the original samples. The graph similarity learning is used to measure the similarity of sample distributions between subspaces with different dimensions. Then multiple DPSSs are obtained. In SDFE module, the distribution characteristics of the samples are described by combining the local responses of the feature space extracted by the sub-RBFNN in each DPSS. The experimental results show that the proposed DPSS-RBFNN with few kernels can achieve higher accuracy in the recognition task than state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

[Display omitted]

Background: This study aimed to investigate staging changes for Chinese breast cancer patients assessed by the 7th(anatomic) and 8th(prognostic) editions of the AJCC staging manual, and to explore the predictive factors for these changes.Methods: Data of patients who received curative surgery for stage I-III breast cancer at Ningbo Medical Center Lihuili Eastern Hospital were retrospectively reviewed. The assessment of staging was according to the criteria of the 7thand 8theditions of the AJCC staging manual. Univariate and multivariate logistic regression analyses were performed to analyze the associations between staging changes and clinicopathological characteristics.Results: Staging changes were found in 59.37% of patients and were more likely to be seen in stage IIIA (96.10%) and IIA (85.94%), then IIB (70.33%), IB (68.75%), followed by IA (36.17%) and IIIC (30.08%). In univariate analysis, staging changes were associated with tumor location, clinical tumor size, clinical axillary lymph node status and Ki67 index. However, multivariate analysis found that staging changes were significantly associated with tumor size >2 cm (odds ratio [OR] = 3.263, 95% confidence interval [95% CI], 2.638-4.036), lymph node involvement (OR = 2.261, 95% CI, 1.830-2.794) and high Ki-67 index (OR = 1.661, 95% CI 1.343-2.054).Conclusions: Our study demonstrated that there were marked staging changes when 2 different editions of the AJCC staging manual were used. Since prognostic biomarkers are available in routine clinical practice, the more recent staging manual should be followed to select better systemic therapy and give better outcomes for Chinese breast cancer patients.

To report a case of da Vinci robotic-assisted radical resection of hilar cholangiocarcinoma. This case is shared to discuss the techniques, methods and ideas for the treatment of hilar cholangiocarcinoma.

The development of ultrafast detectors for electron microscopy (EM) opens a new door to exploring dynamics of nanomaterials; however, it raises grand challenges for big data processing and storage. Here, we combine deep learning and temporal compressive sensing (TCS) to propose a novel EM big data compression strategy. Specifically, TCS is employed to compress sequential EM images into a single compressed measurement; an end-to-end deep learning network is leveraged to reconstruct the original images. Owing to the significantly improved compression efficiency and built-in denoising capability of the deep learning framework over conventional JPEG compression, compressed videos with a compression ratio of up to 30 can be reconstructed with high fidelity. Using this approach, considerable encoding power, memory, and transmission bandwidth can be saved, allowing it to be deployed to existing detectors. We anticipate the proposed technique will have far-reaching applications in edge computing for EM and other imaging techniques.

This paper presents a two dimensional two-way coupled model combining Smoothed Particle Hydrodynamics (SPH) based on the Navier–Stokes equations (NSE) and OceanWave3D based on the fully nonlinear potential flow theory (FNPT) in order to efficiently simulate non-linear waves and wave–structure interaction problems. The two models are strongly coupled in space and time domains using a fixed overlapping zone, wherein the information from both solvers is exchanged by relaxation functions. In the SPH model, an open relaxation boundary, which is implemented as open and relaxation zones, is used in the coupling region. Horizontal velocity and free surface elevation in the open and relaxation zones are obtained from OceanWave3D, while vertical velocity and density in the open zones are interpolated from the relaxation region. OceanWave3D requires the free surface elevation and vertical velocity at the free surface from SPH in the coupled region. The coupled model is tested by modelling a regular wave, irregular wave and wave over a submerged bar and an oscillating water column (OWC) device. The results demonstrate that the coupled model can produce satisfactory results with less computational time than the SPH-only model. • A two-way coupling model between OceanWave3D and SPH is proposed. • Coupling region in SPH is implemented as open relaxation boundary. • The coupled model is computationally cheaper than an SPH-only model. [ABSTRACT FROM AUTHOR]

This paper presents a new Smoothed Particle Hydrodynamics (SPH) parallel framework, which is designed for free surface flows and is scalable on a High Performance Computer (HPC). The framework is accomplished by adopting a Message Passing Interface (MPI) approach with a domain partitioning strategy. A regular background grid is used to partition the entire computational domain and each subdomain is labelled using an index ordering method. Adjacent subdomains can be determined by the index list, and avoid global communications in the particle distribution process. Within the local grid, the grid is divided into an internal grid as well as an interactive grid to identify the particles for which information is to be transferred. The implementation of the dynamic loading balance strategy considers two different ways of determining loading: computation particle numbers and running time. The dynamic load balance strategy repositions neighbouring subdomains based on the local load imbalance between cores. To demonstrate the framework's capacity and distinctive properties, a variety of free surface flow benchmarks are studied. Intensive numerical experiments at various scales are used to assess the performance in detail. • We propose a new MPI-based parallel SPH framework for free-surface flow. • A dynamic load balance strategy is applied in the parallel SPH framework. • The SPH framework works well for free surface conditions with large deformation. • Time cost-based balancing strategy is more efficient than the particle number-based. [ABSTRACT FROM AUTHOR]