Explore the vast range of titles available.

This study aims to detect seeded porosity during metal additive manufacturing by employing convolutional neural networks (CNN). The study demonstrates the application of machine learning (ML) in in-process monitoring. Laser powder bed fusion (LPBF) is a selective laser melting technique used to build complex 3D parts. The current monitoring system in LPBF is inadequate to produce safety-critical parts due to the lack of automated processing of collected data. To assess the efficacy of applying ML to defect detection in LPBF by in-process images, a range of synthetic defects have been designed into cylindrical artefacts to mimic porosity occurring in different locations, shapes, and sizes. Empirical analysis has revealed the importance of accurate labelling strategies required for data-driven solutions. We formulated two labelling strategies based on the computer-aided design (CAD) file and X-ray computed tomography (XCT) scan data. A novel CNN was trained from scratch and optimised by selecting the best values of an extensive range of hyper-parameters by employing a Hyperband tuner. The model’s accuracy was 90% when trained using CAD-assisted labelling and 97% when using XCT-assisted labelling. The model successfully spotted pores as small as 0.2mm. Experiments revealed that balancing the data set improved the model’s precision from 89% to 97% and recall from 85% to 97% compared to training on an imbalanced data set. We firmly believe that the proposed model would significantly reduce post-processing costs and provide a better base model network for transfer learning of future ML models aimed at LPBF micro-defects detection.

The advancement of single cell RNA-sequencing (scRNA-seq) technology has enabled the direct inference of co-expressions in specific cell types, facilitating our understanding of cell-type-specific biological functions. For this task, the high sequencing depth variations and measurement errors in scRNA-seq data present two significant challenges, and they have not been adequately addressed by existing methods. We propose a statistical approach, CS-CORE, for estimating and testing cell-type-specific co-expressions, that explicitly models sequencing depth variations and measurement errors in scRNA-seq data. Systematic evaluations show that most existing methods suffered from inflated false positives as well as biased co-expression estimates and clustering analysis, whereas CS-CORE gave accurate estimates in these experiments. When applied to scRNA-seq data from postmortem brain samples from Alzheimer’s disease patients/controls and blood samples from COVID-19 patients/controls, CS-CORE identified cell-type-specific co-expressions and differential co-expressions that were more reproducible and/or more enriched for relevant biological pathways than those inferred from existing methods. Inferring co-expressions with scRNA-seq data is challenging, and existing methods suffer from inflated false positives and biases. Here, the authors proposed CS-CORE, which yields unbiased estimates and identifies co-expressions that are more reproducible and biologically relevant for scRNA-seq data.

In this era of information explosion, recommendation systems play a key role in helping users to uncover content of interest among massive amounts of information. Pursuing a breadth of recall while maintaining accuracy is a core challenge for current recommendation systems. In this paper, we propose a new recommendation algorithm model, the interactive higher-order dual tower (IHDT), which improves current models by adding interactivity and higher-order feature learning between the dual tower neural networks. A heterogeneous graph is constructed containing different types of nodes, such as users, items, and attributes, extracting richer feature representations through meta-paths. To achieve feature interaction, an interactive learning mechanism is introduced to inject relevant features between the user and project towers. Additionally, this method utilizes graph convolutional networks for higher-order feature learning, pooling the node embeddings of the twin towers to obtain enhanced end-user and item representations. IHDT was evaluated on the MovieLens dataset and outperformed multiple baseline methods. Ablation experiments verified the contribution of interactive learning and high-order GCN components.

In anomaly detection tasks, labeled defect data are often scarce. Unsupervised learning leverages only normal samples during training, making it particularly suitable for anomaly detection tasks. Among unsupervised methods, normalizing flow models have shown distinct advantages. They allow precise modeling of data distributions and enable direct computation of sample log-likelihoods. Recent work has largely focused on feature fusion strategies. However, most of the flow-based methods emphasize spatial information while neglecting the critical role of channel-wise features. To address this limitation, we propose GCAFlow, a novel flow-based model enhanced with a global context-aware channel attention mechanism. In addition, we design a hierarchical convolutional subnetwork to improve the probabilistic modeling capacity of the flow-based framework. This subnetwork supports more accurate estimation of data likelihoods and enhances anomaly detection performance. We evaluate GCAFlow on three benchmark anomaly detection datasets, and the results demonstrate that it consistently outperforms existing flow-based models in both accuracy and robustness. In particular, on the VisA dataset, GCAFlow achieves an image-level AUROC of 98.2% and a pixel-level AUROC of 99.0%.

Pipeline leakage can lead to catastrophic consequences, and traditional sensor-based detection methods often struggle to identify changes caused by slow or minor leaks. This paper proposes a real-time machine vision-based method for detecting liquid leakage in pipelines, suitable for complex industrial scenarios. By extracting droplet foreground regions and constructing a detection model based on the contour and motion features of droplets, the proposed method effectively filters out interference from lighting variations, equipment vibrations, and personnel movement in industrial environments, while accurately identifying the vertical motion characteristics of dripping liquids. An experimental platform was established to validate the effectiveness of the proposed approach. The results demonstrate that the proposed method achieves a detection rate of 98.04%, a false alarm rate of 5.26%, and a processing speed of 90.71 fps. Comparative experiments show that this method significantly outperforms traditional approaches, such as the dense optical flow method, which yields a higher false alarm rate and a processing speed of only 2.2 fps under the same test conditions. These findings confirm that our approach offers a more accurate and efficient solution for real-time pipeline liquid leakage detection.

Zero-dimensional (0D)–one-dimensional (1D) mixed-dimensional van der Waals (MvdW) heterostructures have shown great potential in electronic/optoelectronic applications. However, addressing the interface barrier modulation and charge-transfer mechanisms remain challenging. Here, we develop an analytic model to illustrate the open-circuit voltage and charge-transfer state energy in PbS ₓ Se ₁− ₓ -quantum dots (QDs)/MoS ₂ -nanotube (NT) 0D–1D MvdW heterostructures based on atomic-bond-relaxation approach, Marcus theory and modified-detailed balance principle. We find that the band alignment of PbS ₓ Se ₁− ₓ -QDs/MoS ₂ -NT heterostructures undergoes a transition from type II to type I, and the threshold of size is around 5.6 nm for x = 1, which makes the system suitable for various devices including photocatalytic device, light-emission device and solar cell under different sizes. Our results not only clarify the underlying mechanism of interfacial charge-transfer in the heterostructures, but also provide unique insight and new strategy for designing multifunctional and high-performance 0D–1D MvdW heterostructure devices.

Epigenetics including DNA and RNA modifications have always been the hotspot field of life sciences in the post-genome era. Since the first mapping of N6-methyladenosine (m 6 A) and the discovery of its widespread presence in mRNA, there are at least 160-170 RNA modifications have been discovered. These methylations occur in different RNA types, and their distribution is species-specific. 5-methylcytosine (m 5 C) has been found in mRNA, rRNA and tRNA of representative organisms from all kinds of species. As reversible epigenetic modifications, m 5 C modifications of RNA affect the fate of the modified RNA molecules and play important roles in various biological processes including RNA stability control, protein synthesis, and transcriptional regulation. Furthermore, accumulative evidence also implicates the role of RNA m 5 C in tumorigenesis. Here, we review the latest progresses in the biological roles of m 5 C modifications and how it is regulated by corresponding “writers”, “readers” and “erasers” proteins, as well as the potential molecular mechanism in tumorigenesis and cancer immunotherapy.

Anisotropic effect of two-dimensional materials remains one of the most attractive properties, introducing an additional degree of freedom for tuning physical performances. In this work, we investigate the anisotropies of electronic, transport, piezoelectric, and optoelectronic properties of Janus In _2 XY ( X / Y = S, Se, Te) monolayers (J-In _2 XY MLs) by performing first-principles calculations. We find that such J-In _2 XY MLs possess moderate bandgap (2.07–2.29 eV), high carrier mobility (∼10 ^3 cm ^−2 V ^−1 s ^−1 ), visible light absorption (∼10 ^5 cm ^−1 ), large out-of-plane piezoelectric response$({d_{31}} = 0.35{\\text{ pm}}\\;{{\\text{V}}^{ - 1}}{\\text{)}}$and ultra-soft mechanical nature$({C_{11}} = 32.84{\\text{ N}}\\;{{\\text{m}}^{ - 1}}{\\text{)}}$. We construct a kind of J-In _2 XY- based phototransistor to investigate the optoelectronic properties under linearly polarized light. We find that the low recombination probability of photogenerated carriers ensured by anisotropic effect enhances the photocatalytic potential of J-In _2 XY MLs. And the pivotal role induced by anisotropy in photocurrent can cause a prominent on/off ratio (∼100), considerable responsivity (0.038 AW ^−1 , 0.036 AW ^−1 ) and external quantum efficiency (10.6%, 11.5%). Our study provides an avenue for the design of future anisotropic J-In _2 XY- based multifunctional optoelectronic device.

•Heritability and correlation estimation for 33 anatomically defined regions.•Network and cluster analysis for genetic correlation characterize human brain's functional architecture.•Composite-linkage disequilibrium score regression reveals correlations between composite phenotypes and complex traits. Heritability and genetic covariance/correlation quantify the marginal and shared genetic effects across traits. They offer insights on the genetic architecture of complex traits and diseases. To explore how genetic variations contribute to brain function variations, we estimated heritability and genetic correlation across cortical thickness, surface area, and volume of 33 anatomically predefined regions in left and right hemispheres, using summary statistics of genome-wide association analyses of 31,968 participants in the UK Biobank. To characterize the relationships between these regions of interest, we constructed a genetic network for these regions using recursive two-way cut-offs in similarity matrices defined by genetic correlations. The inferred genetic network matches the brain lobe mapping more closely than the network inferred from phenotypic similarities. We further studied the associations between the genetic network for brain regions and 30 complex traits through a novel composite-linkage disequilibrium score regression method. We identified seven significant pairs, which offer insights on the genetic basis for regions of interest mediated by cortical measures.

Basaltic eruptions are the most common form of volcanism on Earth and planetary bodies. The low viscosity of basaltic magmas inhibits fragmentation, which favours effusive and lava-fountaining activity, yet highly explosive, hazardous basaltic eruptions occur. The processes that promote fragmentation of basaltic magma remain unclear and are subject to debate. Here we used a numerical conduit model to show that a rapid magma ascent during explosive eruptions produces a large undercooling. In situ experiments revealed that undercooling drives exceptionally rapid (in minutes) crystallization, which induces a step change in viscosity that triggers magma fragmentation. The experimentally produced textures are consistent with basaltic Plinian eruption products. We applied a numerical model to investigate basaltic magma fragmentation over a wide parameter space and found that all basaltic volcanoes have the potential to produce highly explosive eruptions. The critical requirements are initial magma temperatures lower than 1,100 °C to reach a syn-eruptive crystal content of over 30 vol%, and thus a magma viscosity around 105 Pa s, which our results suggest is the minimum viscosity required for the fragmentation of fast ascending basaltic magmas. These temperature, crystal content and viscosity requirements reveal how typically effusive basaltic volcanoes can produce unexpected highly explosive and hazardous eruptions.