Comprehensive Representation Research Articles

Nonparametric and Continuous Variable-Based Stratigraphic Modelling from Sparse Boreholes using Signed Distance Function and Bayesian Compressive Sensing

An accurate stochastic interpretation of subsurface stratigraphy with quantified uncertainty can benefit the subsequent risk management of geotechnical infrastructure. Traditional approaches to developing geological cross-sections from sparse boreholes typically require the calibration or definition of empirical model parameters and functions, which may introduce subjectivity and bias. In this study, a non-parametric and continuous variable-based spatial predictor that leverages the signed distance function and Bayesian compressive sensing (BCS) is proposed for subsurface stratigraphic modelling. The proposed method transforms sparse categorical borehole data from a low-dimensional space into continuous variables in a high-dimensional space, enabling a comprehensive representation of more implicit characteristics of intricate geological patterns. This transformation facilitates the use of the continuous-variable-based BCS for non-parametric spatial prediction. The most probable geological cross-section and uncertainty qualification plot are derived after transforming spatially interpreted fields of continuous variables back into soil types. The performance of the proposed method is demonstrated using synthetic and real-world cases. Results indicate that the proposed approach can handle intricate stratigraphic scenarios characterized by complex geological structures, such as crossed-inclined, folded, inclined-folded, and interbedded strata, in a data-driven and non-parametric manner. The advantages of the proposed method over existing spatial predictors for developing geological cross-sections are also demonstrated.

Multimodal Deep Representation Learning for Quantum Cross-Platform Verification.

Cross-platform verification, a critical undertaking in the realm of early-stage quantum computing, endeavors to characterize the similarity of two imperfect quantum devices executing identical algorithms, utilizing minimal measurements. While the random measurement approach has been instrumental in this context, the quasiexponential computational demand with increasing qubit count hurdles its feasibility in large-qubit scenarios. To bridge this knowledge gap, here we introduce an innovative multimodal learning approach, recognizing that the formalism of data in this task embodies two distinct modalities: measurement outcomes and classical description of compiled circuits on explored quantum devices, both containing unique information about the quantum devices. Building upon this insight, we devise a multimodal neural network to independently extract knowledge from these modalities, followed by a fusion operation to create a comprehensive data representation. The learned representation can effectively characterize the similarity between the explored quantum devices when executing new quantum algorithms not present in the training data. We evaluate our proposal on platforms featuring diverse noise models, encompassing system sizes up to 50 qubits. The achieved results demonstrate an improvement of 3 orders of magnitude in prediction accuracy compared to the random measurements and offer compelling evidence of the complementary roles played by each modality in cross-platform verification. These findings pave the way for harnessing the power of multimodal learning to overcome challenges in wider quantum system learning tasks.

Structural mechanism of glass transition uncovered by unsupervised machine learning

Uncovering the structural origins of the ubiquitous dynamic arrest phenomenon at the glass transition has long been a challenge due to the difficulty in identifying a rational structural representation from a disordered medium. To address this challenge, we propose a novel approach based on unsupervised learning to define a set of structural fingerprints. In this approach, complex local atomic environments, ranging from short to medium range, are captured by the discretized radial distribution function and projected onto a simple two-dimensional space using a neural network-based autoencoder. This two-dimensional space is characterized by two static structural indicators, P1 and P2, providing a comprehensive and user-friendly representation of the mysterious “glassy structure”. By employing Gaussian mixture modeling, the structural space is autonomously divided into three sections, each representing a unique cluster with similar environments. These indicators not only elucidate the glass transition but also allow for the quantitative prediction of activation barriers for local structural excitations. Furthermore, the unsupervised clustering technique can distinguish between the structural features of “hard zones” and “soft zones”, as well as recently proposed superfast “liquid-like” atoms in glass. This unsupervised machine learning approach demonstrates the utility of seemingly agnostic local structure in amorphous materials, offering insights into the long-sought structural origins of the glass transition.

A privacy-preserving framework with multi-modal data for cross-domain recommendation

Cross-domain recommendation (CDR) aims to enhance the recommendation accuracy in a target domain with sparse data by leveraging rich information in a source domain, thereby addressing the data-sparsity problem. Some existing CDR methods highlight the advantages of extracting domain-common and domain-specific features to learn comprehensive user and item representations. However, these methods cannot effectively disentangle these components, as they often rely on simple user-item historical interaction information (such as ratings, clicks, and browsing), neglecting the rich multi-modal features. In addition, they do not protect user-sensitive data from potential leakage during knowledge transfer between domains. To address these challenges, we propose a Privacy-Preserving Framework with Multi-Modal Data for Cross-Domain Recommendation, called P2M2-CDR. Specifically, we first design a multi-modal disentangled encoder that utilizes multi-modal information to disentangle more informative domain-common and domain-specific embeddings. Furthermore, we introduce a privacy-preserving decoder to mitigate user privacy leakage during knowledge transfer. Local differential privacy (LDP) is used to obfuscate disentangled embeddings before the inter-domain exchange, thereby enhancing privacy protection. To ensure both consistency and differentiation among these obfuscated disentangled embeddings, we incorporate contrastive learning-based domain-inter and domain-intra losses. Extensive experiments conducted on six CDR tasks from two real-world datasets demonstrate that P2M2-CDR outperforms other state-of-the-art single- and cross-domain baselines. The code is available at https://github.com/Lili1013/P2M2-CDR.

The inertial-based gait normalcy index of dual task cost during turning quantifies gait automaticity improvement in early-stage Parkinson’s rehabilitation

BackgroundThe loss of gait automaticity is a key cause of motor deficits in Parkinson’s disease (PD) patients, even at the early stage of the disease. Action observation training (AOT) shows promise in enhancing gait automaticity. However, effective assessment methods are lacking. We aimed to propose a novel gait normalcy index based on dual task cost (NIDTC) and evaluate its validity and responsiveness for early-stage PD rehabilitation.MethodsThirty early-stage PD patients were recruited and randomly assigned to the AOT or active control (CON) group. The proposed NIDTC during straight walking and turning tasks and clinical scale scores were measured before and after 12 weeks of rehabilitation. The correlations between the NIDTCs and clinical scores were analyzed with Pearson correlation coefficient analysis to evaluate the construct validity. The rehabilitative changes were assessed using repeated-measures ANOVA, while the responsiveness of NIDTC was further compared by t tests.ResultsThe turning-based NIDTC was significantly correlated with multiple clinical scales. Significant group-time interactions were observed for the turning-based NIDTC (F = 4.669, p = 0.042), BBS (F = 6.050, p = 0.022) and PDQ-39 (F = 7.772, p = 0.011) tests. The turning-based NIDTC reflected different rehabilitation effects between the AOT and CON groups, with the largest effect size (p = 0.020, Cohen’s d = 0.933).ConclusionThe turning-based NIDTC exhibited the highest responsiveness for identifying gait automaticity improvement by providing a comprehensive representation of motor ability during dual tasks. It has great potential as a valid measure for early-stage PD diagnosis and rehabilitation assessment.Trial registration Chinese Clinical Trial Registry: ChiCTR2300067657

Multi-Behavior Recommendation with Personalized Directed Acyclic Behavior Graphs

A well-developed recommendation system can not only leverage multi-typed interactions (such as page view, add-to-cart, and purchase) to better identify user preferences, but also demonstrate high performance, low complexity, and strong interpretability. However, many existing solutions for multi-behavior recommendation fall short of intuitive modeling of real-world scenarios, leading to overly complex models with massive parameters and cumbersome components. In particular, they share two critical limitations: (1) Some pioneering models are built upon the strict assumption of cascade effects across behaviors, which contradicts multifarious behavior paths in practical applications. (2) Existing approaches fail to explicitly capture the unique idiosyncrasies of users and even neglect the inherent nature of items involved in the multi-behavior interactions. To this end, we propose a novel Directed Acyclic Graph Convolutional Network (DA-GCN) for the multi-behavior recommendation task. Specifically, we pinpoint the partial order relations within the monotonic behavior chain and extend it to personalized directed acyclic behavior graphs to exploit behavior dependencies. Then, a GCN-based directed edge encoder is employed to distill rich collaborative signals embodied by each directed edge. In light of the information flows over the directed acyclic structure, we propose an attentive aggregation module to gather messages from all potential antecedent behaviors, representing distinct perspectives to understand the terminated behavior. Thus, we obtain comprehensive representations for the follow-up behavior through learnable distributions over its preceding behaviors, explicitly reflecting personalized interactive patterns of users and underlying properties of items simultaneously. Finally, we design a customized multi-task learning objective for flexible joint optimization. Extensive experiments on public benchmarking datasets fully demonstrate the superiority of DA-GCN with significant performance improvement and computational efficiency over a wide range of state-of-the-art methods. Our code is available at https://github.com/xizhu1022/DA-GCN .

VTNet: A multi-domain information fusion model for long-term multi-variate time series forecasting with application in irrigation water level

Time series forecasting is intricately tied to production and life, garnering widespread attention over an extended period. Enhancing the performance of long-term multivariate time series forecasting (MTSF) poses a highly challenging task, as it requires mining complicated and obscure temporal patterns in many aspects. For this reason, this paper proposes a long-term forecasting model based on multi-domain fusion (VTNet) to adaptively capture and refine multi-scale intra- and inter-variate dependencies. In contrast to previous techniques, we devise a dual-stream learning architecture. Firstly, the fast Fourier Transform (FFT) is adopted to extract frequency domain information. The original sequences are then transformed into 2D visual features in the temporal-frequency domain, and a 2D-TBlock is designed for multi-scale dynamic learning. Secondly, a combination of convolution and recurrent networks continues to explore the local temporal features and preserve the global trend. Finally, multi-modal circulant fusion is applied to achieve a more comprehensive and enriched feature fusion representation, further promoting overall performance. Extensive experiments are conducted on 9 public benchmark datasets and the real-world irrigation water level to showcase VTNet’s promoted performance and generalization. Moreover, VTNet yields 46.93% and 25.36% relative improvements for water level forecasting, revealing its potential application value in water-saving planning and extreme event early warning.

Semantic Segmentation-Driven Integration of Point Clouds from Mobile Scanning Platforms in Urban Environments

Precise and complete 3D representations of architectural structures or industrial sites are essential for various applications, including structural monitoring or cadastre. However, acquiring these datasets can be time-consuming, particularly for large objects. Mobile scanning systems offer a solution for such cases. In the case of complex scenes, multiple scanning systems are required to obtain point clouds that can be merged into a comprehensive representation of the object. Merging individual point clouds obtained from different sensors or at different times can be difficult due to discrepancies caused by moving objects or changes in the scene over time, such as seasonal variations in vegetation. In this study, we present the integration of point clouds obtained from two mobile scanning platforms within a built-up area. We utilized a combination of a quadruped robot and an unmanned aerial vehicle (UAV). The PointNet++ network was employed to conduct a semantic segmentation task, enabling the detection of non-ground objects. The experimental tests used the Toronto 3D dataset and DALES for network training. Based on the performance, the model trained on DALES was chosen for further research. The proposed integration algorithm involved semantic segmentation of both point clouds, dividing them into square subregions, and performing subregion selection by checking the emptiness or when both subregions contained points. Parameters such as local density, centroids, coverage, and Euclidean distance were evaluated. Point cloud merging and augmentation enhanced with semantic segmentation and clustering resulted in the exclusion of points associated with these movable objects from the point clouds. The comparative analysis of the method and simple merging was performed based on file size, number of points, mean roughness, and noise estimation. The proposed method provided adequate results with the improvement of point cloud quality indicators.

A new dataset of leaf optical traits to include biophysical parameters in addition to spectral and biochemical assessment

To enable future improvement on current leaf optical property models, more data incorporating a larger range of measured properties is needed. To this end, a dataset was collected to associate spectral measurements (ultraviolet, visible, and near infrared) with biochemical and biophysical properties of leaves. The leaves represented in this dataset were selected to provide a more comprehensive representation of both tree and agricultural species as well as leaves with a wide variety of color (pigment) expression, surface characteristics, and stages in a leaf lifecycle. Extensive data were collected for each of the 290 leaf samples studied in this project including multiple spectral measurement orientations and ranges, biochemical assessment, and biophysical assessment of that has not previously been a focus in other leaf datasets. The methods and results associated with this dataset are described in this work.

A study on improving drug–drug interactions prediction using convolutional neural networks

Appropriate studies on drug–drug interactions (DDIs) can evade possible adverse side effects due to the ingestion of multiple drugs. This paper proposes a novel framework called Similarity Network Fusion and Hybrid Convolutional Neural Network (SNF–HCNN) to predict the DDIs better. The proposed framework leverages data from DrugBank, PubChem, and SIDER. Seven critical drug features are extracted: Target, Transporter, Enzymes, Chemical substructure, Carrier, Offside, and Side effects. The Jaccard Similarity measure evaluates the similarity of drug features to construct a comprehensive similarity matrix that effectively captures potential drug relationships and patterns. The similarity selection process identifies the most relevant features, reduces redundancy, and enhances identifying potential drug interactions. Integrating Similarity Network Fusion (SNF) with the selected similarity matrix ensures a comprehensive representation of drug features and leads to superior accuracy compared to conventional methods. Our experimental results demonstrate the effectiveness of the proposed hybrid convolutional neural network (HCNN) architectures, such as CNN+LR (CNN+Logistic Regression), CNN+RF (CNN+Random Forest), and CNN+SVM (CNN+Support Vector Machine), showing impressive accuracies of 95.19%, 94.45%, and 93.65%, respectively. Moreover, CNN+LR outperforms other approaches regarding precision, sensitivity, F1-score, and AUC score, which implicate better outcomes for ensuring medication safety aspects in clinical settings in the future.

Imaging of Volatile Organic Compounds Using a Single Nanowire-Based Electronic Nose for Future Biomedical Applications.

This study introduces an array of semiconductor oxide single nanowires fabricated using advanced semiconductor processing techniques, including electron beam lithography and thin-film deposition, which is well-suited for large-scale nanowire integration. A four-channel nanowire array consisting of tin oxide (SnO2), indium oxide (In2O3), ferric oxide (Fe3O4), and titanium oxide (TiO2) was developed. As a proof of concept, we converted the response curves of the sensor array to heat maps, enabling comprehensive feature representation. The fabricated electronic nose (E-nose) was utilized to detect three types of volatile organic compounds (VOCs), with the results visualized in a heat map format. Additionally, the performance of each individual sensor was quantitatively studied, highlighting the array's potential for enhanced gas detection and analysis. To further illustrate the interaction between gas molecules and the nanowires, we visualized the gas response results by mapping the sensor's signal changes. These visualizations provide a clear representation of how different gas molecules interact with specific nanowires. For example, the heat maps reveal distinct response patterns for each type of VOC, allowing for the identification and differentiation of gases based on their unique signatures. This visualization technique not only enhances the understanding of gas-nanowire interactions but also demonstrates the effectiveness of the E-nose in distinguishing between various VOCs. The SnO2 nanowire gas sensor showed enhanced gas response compared to other materials. The SnO2 and TiO2 gas sensors showed enhanced response (62 and 56 s) and recovery times (100 and 37 s).

Decoding stakeholder priorities of safety culture preferences in the oil and gas industry

Safety culture is a critical determinant of organisational performance, particularly in high-risk industries especially in oil and gas. Understanding stakeholder preferences is essential for developing effective strategies that enhance safety culture. This study utilised the Analytic Hierarchy Process (AHP) to prioritise stakeholder preferences, identifying key elements of safety culture in Malaysia's oil and gas sector. This study employed a structured methodology to evaluate safety culture within the oil and gas industry, focusing on 18 sub-elements across three key domains: psychological, behavioural, and situational factors. A diverse sample of industry experts was recruited using purposeful and snowball sampling to ensure a comprehensive representation of stakeholder views. The AHP framework was applied to analyse the data, utilizing structured questionnaires and multicriteria decision-making techniques to prioritize the identified safety culture elements. The AHP analysis identified distinct priorities among different professional groups within the oil and gas sector. Safety and Health Practitioners emphasized practical elements such as safety rules and management commitment, while academicians prioritized knowledge and training. Management personnel highlighted the importance of safety ownership and communication, whereas policymakers focused on broader, policy-oriented aspects. The findings suggest that safety culture improvement initiatives should be tailored to address the specific needs and priorities of each professional group. A nuanced understanding of stakeholder preferences is crucial for developing comprehensive strategies that integrate observable behaviours, situational conditions, and psychological factors, ultimately fostering a robust safety culture in the oil and gas industry.

Visualization-based comprehensive feature representation with improved EfficientNet for malicious file and variant recognition

Malicious file attacks seriously affect network and data security, and recognizing malicious files and variants is crucial for preventing network attacks. Faced with the challenge of traditional methods in quickly, effectively, and efficiently recognizing malicious files or variants, visualization-based feature representation methods have shown promising results. However, practical applications encounter issues such as loss of crucial information, high spatiotemporal overhead, and the need for model performance improvement. Therefore, this paper introduces a novel recognition framework focusing on feature representation and model performance. The framework uses the proposed visualization-based comprehensive feature representation method (VCFR) to extract file information into the Gray-Level Co-occurrence Matrix (GLCM), 2-gram frequency matrix, and interval 2-gram frequency matrix, followed by feature fusion to generate the three-channel RGB images. Subsequently, the proposed lightweight model is applied for recognizing those files, which utilizes ideas such as group convolution, channel shuffle, and attention mechanisms to improve model performance while significantly reducing model parameters, size, and FLOPs. In summary, through a series of experiments conducted on manually collected malicious file dataset (MFD) and public dataset MMCC, the proposed framework significantly outperformed other state-of-the-art technologies and has F1-Score as high as 94.10% and 98.58%, respectively, further verifying its outstanding effectiveness and efficiency.

Short-term power load forecasting based on hybrid feature extraction and parallel BiLSTM network

Electricity is a vital resource for societal and economic activities. Accurate electricity load forecasting can effectively reduce costs and enhance energy efficiency. Nevertheless, current research has limitations in effectively extracting and utilizing load forecasting information. This encompasses both undiscovered latent features and models that have not effectively captured long-term dependencies. To address these issues, this paper proposes a novel network architecture, which utilizes a hybrid feature extraction strategy and integrates machine learning methods with deep learning techniques. This method employs a hybrid feature extraction strategy composed of gradient-boosted regression trees, Fisher Score, and mutual information to achieve comprehensive feature representation. The network architecture utilizes a multi-head convolutional neural network and a bi-directional long short-term memory (BiLSTM) with lag parameters. This architecture can simultaneously learn multiple levels and aspects of features, further enhancing the BiLSTM model’s ability to capture long-term dependencies. Through multiple experiments conducted on datasets from Maine, Singapore, New South Wales, Australia, and European countries, our method outperformed the comparison models in three out of the five datasets.

Agritourism as a strategy of remote island’s development: the case study of Kasos

In rural, depopulated areas, agritourism, as a development model, could serve the goal of non- exclusively seasonal tourism while creating additional economic opportunities for the locals who are mainly engaged in agriculture. Kasos is an isolated island of the Dodecanese, Greece, preserving interesting elements of the authentic agricultural practices and architecture like terraces serving the island’s agricultural cultivation, mills and mitata, small stone- built creameries. Kasos served as a case study of AEI program (Sustainable Development of Less Developed Regions and Isolated Areas by Creating New Touristic Resources and Products through Analysis, Documentation and Modelling of Cultural Assets using Innovative ICT Applications, project code:T2EDK- 01278). For thιs island that lies within the wider region of Rhodes, it is possible with the appropriate development of new touristic resources and products through the advancement of the unique architectural, cultural and environmental assets to orient towards them special parts of the touristic flow from Rhodes, creating new touristic activity by increasing the residence beyond the daily or short time tourism, as it stands for now. For the analysis of the study area, research in local archives and communication with the municipality and local stakeholders was conducted. Thus, both the special features of the island and the needs of the local community were identified, organized, classified and presented in GIS maps. The implementation of a GIS for Kasos Island aims to provide a comprehensive representation of the different aspects of the island, providing useful information for policy-making and the development of strategic cultural and agrotourism plans to the local authorities and stakeholders. According to the integrated plan for Kasos by AEI program, the island could preserve its special rural identity while revealing its cultural and environmental assets and attracting external economies within an agritourism model. In this work, we present AEI’s integrated plan for Kasos that refers to the island’s agricultural character, to the existing trail network as well as to the restoration and adaptive reuse of important architectural assets that could serve to the overall revealing of Kasos history and cultural reserve.

Divergence-guided disentanglement of view-common and view-unique representations for multi-view data

In the field of multi-view learning (MVL), it is crucial to extract both common (consistent) and unique (complementary) information across different views. While the focus has traditionally been on acquiring common information, there has been a recent shift towards exploring unique information as well. However, developing an MVL model that can simultaneously capture both common and unique information, thereby facilitating a comprehensive understanding of multi-view data, remains a significant challenge. To address this, we propose the Divergence-guided Multi-view Learning framework (DG-MVL), inspired by information-theoretic learning theory, specifically the generalized divergence measure. This framework employs multi-view autoencoders to disentangle the features obtained from each view into coarse common and unique components. By minimizing the divergence between the coarse common features learned from the common encoder of each view and maximizing the divergence between the coarse unique features from unique encoders simultaneously, we optimize the extraction of both common and unique information. Subsequently, we merge these features to generate a comprehensive and concise representation of the multi-view data, which can be easily utilized for various downstream tasks. We validate our framework on a synthetic multi-view dataset, demonstrating its effectiveness in disentangling common and unique information. Further experiments on various real-world datasets confirm the effectiveness of DG-MVL in capturing common and unique information from multi-view data, resulting in superior classification performance compared to existing methods. Code is available at https://github.com/LMFLRB/DG-MVL.git.

Exploring ncRNA-Drug Sensitivity Associations via Graph Contrastive Learning.

Increasing evidence has shown that noncoding RNAs (ncRNAs) can affect drug efficiency by modulating drug sensitivity genes. Exploring the association between ncRNAs and drug sensitivity is essential for drug discovery and disease prevention. However, traditional biological experiments for identifying ncRNA-drug sensitivity associations are time-consuming and laborious. In this study, we develop a novel graph contrastive learning approach named NDSGCL to predict ncRNA-drug sensitivity. NDSGCL uses graph convolutional networks to learn feature representations of ncRNAs and drugs in ncRNA-drug bipartite graphs. It integrates local structural neighbours and global semantic neighbours to learn a more comprehensive representation by contrastive learning. Specifically, the local structural neighbours aim to capture the higher-order relationship in the ncRNA-drug graph, while the global semantic neighbours are defined based on semantic clusters of the graph that can alleviate the impact of data sparsity. The experimental results show that NDSGCL outperforms basic graph convolutional network methods, existing contrastive learning methods, and state-of-the-art prediction methods. Visualization experiments show that the contrastive objectives of local structural neighbours and global semantic neighbours play a significant role in contrastive learning. Case studies on two drugs show that NDSGCL is an effective tool for predicting ncRNA-drug sensitivity associations.

MolMVC: Enhancing molecular representations for drug-related tasks through multi-view contrastive learning.

Effective molecular representation is critical in drug development. The complex nature of molecules demands comprehensive multi-view representations, considering 1D, 2D, and 3D aspects, to capture diverse perspectives. Obtaining representations that encompass these varied structures is crucial for a holistic understanding of molecules in drug-related contexts. In this study, we introduce an innovative multi-view contrastive learning framework for molecular representation, denoted as MolMVC. Initially, we use a Transformer encoder to capture 1D sequence information and a Graph Transformer to encode the intricate 2D and 3D structural details of molecules. Our approach incorporates a novel attention-guided augmentation scheme, leveraging prior knowledge to create positive samples tailored to different molecular data views. To align multi-view molecular positive samples effectively in latent space, we introduce an adaptive multi-view contrastive loss (AMCLoss). In particular, we calculate AMCLoss at various levels within the model to effectively capture the hierarchical nature of the molecular information. Eventually, we pre-train the encoders via minimizing AMCLoss to obtain the molecular representation, which can be used for various down-stream tasks. In our experiments, we evaluate the performance of our MolMVC on multiple tasks, including molecular property prediction (MPP), drug-target binding affinity (DTA) prediction and cancer drug response (CDR) prediction. The results demonstrate that the molecular representation learned by our MolMVC can enhance the predictive accuracy on these tasks and also reduce the computational costs. Furthermore, we showcase MolMVC's efficacy in drug repositioning across a spectrum of drug-related applications. The code and pre-trained model are publicly available at https://github.com/Hhhzj-7/MolMVC.

FBENet: Feature-Level Boosting Ensemble Network for Hashimoto's Thyroiditis Ultrasound Image Classification.

Distinguishing Hashimoto's thyroiditis (HT) lesions from ordinary thyroid tissues is difficult with ultrasound images. Challenges in achieving high performance of HT ultrasound image classification include the low resolution, blurred features and large area of irrelevant noise. To address these problems, we propose a Feature-level Boosting Ensemble Network (FBENet) for HT ultrasound image classification. Specifically, to capture the features of suspicious HT lesions efficiently, an Ensemble Feature Boosting Module (EFBM) is introduced into the feature-level ensemble to boost the blurred features. Then, the spatial attention mechanism is adopted in backbone models to improve the feature focusing performance and representation ability. Furthermore, feature-level ensemble technique is employed in the training process to achieve more comprehensive feature representation ability. Experimentally, FBENet was trained on 6,503 HT ultrasound images, and tested on 1,626 HT ultrasound images with 82.92% accuracy and 89.24% AUC on average.

A Combined Scheme based on the Multiscale Stochastic Perturbed Parameterization Tendencies and Perturbed Boundary Layer Parameterization for a Global Ensemble Prediction System

Abstract Stochastic representations of model uncertainties are of great importance for the performance of ensemble prediction systems (EPSs). The stochastically perturbed parameterization tendencies (SPPT) scheme with a single-scale random pattern has been used in the operational global EPS of China Meteorological Administration (CMA-GEPS) since 2018. To deal with deficiencies in this operational single-scale SPPT scheme, a combined scheme based on the multiscale SPPT (mSPPT) scheme and the stochastically perturbed parameterization for the planetary boundary layer (SPP-PBL) scheme is developed. In the combined scheme, the mSPPT component aims to expand model uncertainties characterized by SPPT at mesoscale, synoptic scale, and planetary scale. The SPP-PBL component with six vital parameters is used to capture uncertainties in PBL processes, which is underrepresented by SPPT for the tapering treatment within PBL. Comparisons between the operational SPPT scheme and the mSPPT scheme reveal that the mSPPT scheme can generate more improvements in both ensemble reliability and forecast skills mainly in tropics. Besides, additional benefits from SPP-PBL on top of mSPPT are shown to be primarily distributed in tropics at the lower layers below 850 hPa and surface. Furthermore, the combined scheme of mSPPT and SPP-PBL is suggested to yield better spread–error relationships and forecast skills than the operational SPPT scheme in terms of objective verification scores for standard upper-air variables and surface parameters. A case study for the extreme precipitation event on 20 July 2021 in Henan Province of China also demonstrates the better ability of the combined scheme in forecasting the precipitation intensity and location. Significance Statement A comprehensive and reasonable representation of model uncertainties helps to improve the performance of ensemble prediction systems (EPSs). Despite the popular usage in simulating model uncertainties, the stochastically perturbed parameterization tendencies (SPPT) scheme possesses several shortcomings. To overcome this, a combined scheme based on the multiscale SPPT (mSPPT) scheme and the stochastically perturbed parameterization for the planetary boundary layer (SPP-PBL) scheme is proposed. Based on the global EPS of China Meteorological Administration (CMA-GEPS), additional benefits from the independent usage of mSPPT and SPP-PBL are disclosed. And the combined scheme can inherit the merits of mSPPT and SPP-PBL and generate more improvements on ensemble performance than the original single-scale SPPT scheme. This research provides a guideline for future upgradation of CMA-GEPS.