Prioritized Experience Replay enables the model to learn more about relatively important samples by artificially changing their accessed frequencies. However, this non-uniform sampling method shifts the state-action distribution that is originally used to estimate Q-value functions, which brings about the estimation deviation. In this article, a novel off-policy reinforcement learning training framework

Deep learning (DL) applications have potential for improving the accuracy of type II diabetes diagnoses. However, existing DL applications for the diagnosis of type II diabetes have several drawbacks. For example, they maximize overall diagnostic performance rather than the diagnostic performance for each patient, they do not use objective rules to identify whether a patient has type II diabetes, and

Image restoration is a challenging and complex problem involving recovering the original clear image from a degraded or noisy image. In the medical field, image restoration techniques can significantly improve the quality of endoscopic images, helping doctors make more accurate diagnoses and providing higher-quality data support for computer vision-assisted detection. Existing methods for image restoration

The generation of electrical energy using piezoelectric devices represents a promising alternative due to the high charge density these materials can generate. Cantilever beam devices modeled using finite element methods are commonly used in studies focused on the conversion of mechanical energy into electrical energy. With this, the influence of specific variables and parameters can be analyzed through

Rapid assessment of building damage levels has become very important and has received considerable attention in structural engineering. Traditional methods for this work involve manual inspection, which is often tedious and time-consuming. Deep learning technology in computer vision has developed rapidly in recent years and has proven its superiority. This paper aims to develop an efficient approach

Recently, the prosperity of linear models has raised questions about capturing the sequential capabilities of Transformer forecasters. Although the latest Transformer-based studies have alleviated some of these concerns, the limited information utilization still constrains the model’s comprehensive exploration of complex dependencies, as these forecasters often prioritize global dependence on time

The vectorized representation of a knowledge graph is essential for effectively utilizing its implicit knowledge. Graph neural networks (GNNs) are particularly adept at learning graph representations due to their ability to handle graph topologies. However, GNN-based approaches face two main challenges: first, they fail to differentiate between the types of adjacent nodes during the information aggregation

In this paper, concrete cracking is investigated in dynamics through finite element modeling. A probabilistic approach is employed to translate the effects of material heterogeneity on tensile strength and fracture energy. Both parameters depend on compressive strength and heterogeneity degree (volumetric ratio between finite element and largest aggregate). Material softening is modeled through damage

Many soft biological structures have natural features of viscoelastic and hyperelastic materials. Research focused on the growth biomechanics of these structures is challenging from theoretical and experimental points of view, especially when irregular forms/defects of biological objects should be considered. To this aim, an effort is made in this paper to develop a general nonlinear finite element

The Bayesian model updating method usually involves tens of thousands of finite element model calculations, which will bring huge computational costs to large structures such as bridges. To reduce the computational costs, this paper develops a highly efficient Bayesian model updating method based on a new static homotopy surrogate model. The new surrogate model is established on the basis of the finite

Surface quality is critical for the performance of high-end equipment, with defects potentially leading to severe operational failures. Current defect detection methods face challenges: 2D imaging lacks the ability to capture scratch depth, limiting quantitative damage assessment, while 3D point cloud methods are costly and time-consuming, hindering scalability in manufacturing. This study proposes

Graph data-driven machine learning methods for human activity recognition (HAR) have achieved success recently using sufficient data. In the realm of everyday life, we encounter a notable challenge: the scarcity of labeled radar samples. This limitation is compounded by the stark disparities in data distribution between simulated and measured activity domains. In this article, a generalized graph contrastive

The accurate determination of results in decision analysis is usually predicated on the association between two factors. Although generating data for analytical purposes presents an apparent hurdle, the data obtained may present hurdles in its interpretation. Correlation coefficients can be used to analyze the interaction between two factors and their variations. These coefficients deliver an objective

Ischemic stroke, a leading cause of death and disability, necessitates accurate detection and automatic segmentation of lesions. While diffusion weight imaging is crucial, its single modality limits the detection of subtle lesions and artifacts. To address this, we propose a multi-modality, multi-level fusion network (M2FNet) that aggregates salient features from different modalities across various

Homomorphic encryption is a recent and fundamental breakthrough in modern cryptography, which allows the performance of operations on encrypted data without unveiling the data. Leveraging quantum mechanics principles, quantum computers can potentially solve certain computational problems exponentially faster than classical computers. This immense computational power offers new possibilities for various

This paper proposes a novel data-driven surrogate model for predicting the hysteresis response, i.e., axial force – axial deformation, of steel braces in concentrically braced frames under seismic loading using transfer learning-based artificial neural networks. Transfer learning is utilized to leverage pre-trained baseline long short-term memory networks and transfer its knowledge to the new hysteresis

Cross-domain sequential recommendation (CDSR) aims to predict future sequential interactions in a target domain by analyzing historical sequence data from different domains. A significant challenge in CDSR is the accurate capture of user preferences based on the target domain and multiple domains. Existing methodologies to enhance the performance of the target domain primarily focus on learning preferences

Predicting the Remaining Useful Life (RUL) of industrial equipment is essential for proactive maintenance and health assessment, particularly under the computational constraints of edge devices. While deep learning methods, such as Long Short-Term Memory (LSTM) networks, excel at modeling complex time series, their high computational cost often restricts real-time deployment. To address this challenge

Unsupervised feature selection (UFS) methods play a crucial role in improving the efficiency of extracting relevant information and reducing computational complexity in the context of big data analysis. Despite notable advancements in the field of unsupervised feature selection for large-scale datasets, many UFS methods still remain redundant and irrelevant features during the feature selection process

Traditional recommendation systems tend to focus on accuracy and prefer recommending popular items, resulting in non-popular items rarely being exposed to users. However, recommending non-popular items to enhance users’ novelty experience is also crucial. Currently, many researchers are dedicated to multi-objective recommendation studies. Nevertheless, existing multi-objective recommendation algorithms

Unsupervised multi-source domain adaptation methods transfer knowledge learned from multiple labeled source domains to an unlabeled target domain. Existing methods assume that all source domain data can be accessed directly. However, such an assumption is unrealistic and causes data privacy concerns, especially when the source domain labels include personal information. In such a setting, it is prohibited

Decentralized federated learning (DFL) represents a distributed learning framework where participating nodes independently train local models and exchange model updates with proximate peers, circumventing the reliance on a centralized orchestrator. This paradigm effectively mitigates server-induced bottlenecks and eliminates single points of failure, which are inherent limitations of centralized federated

The impact of airborne pollutants present in the environment, entering the body through breathing, can cause significant risks of respiratory and heart-related health problems for individuals. For this, different air purifiers are commonly used to eliminate delicate particulate matter PM2.5, and various studies have examined their effectiveness. This paper aims to analyze airborne pollutants and, by

Label correlations, as important prior information, are essential to enhance the classification performance in Multi-Instance Multi-Label (MIML) algorithms, but existing models always leverage global label correlations which are less informative. Furthermore, classifier optimization is also crucial for MIML classification results, previous works do not frequently seek to optimize multi objectives simultaneously

A model of train–track–bridge coupled system is proposed to study the interactions between structures in greater detail. The new model employs a meshless method to numerically simulate the box girder bridge and track slab. In the dynamic analysis, the system at each time step is abstracted into a graph structure and trained using a graph neural network to develop a surrogate prediction model. The graph

The melt-cast charging process, widely used in warheads for its adaptability, cost efficiency, and automation, requires optimization to minimize defects such as shrinkage cavities and porosity that compromise explosive quality, destructive power, and safety, particularly in large-volume munitions. The hot mandrel technique, by providing localized heating during solidification, helps maintain an open

A new method to compute the dynamic response of periodic waveguides with localised nonlinearities is introduced and used to investigate the nonlinear shift of a band-edge mode in the bandgap of a locally resonant phononic structure. This nonlinear extension of the Wave Finite Element Method (WFEM) uses a finite-element discretisation of arbitrarily complex unit-cells, and leverages Floquet–Bloch theory

At present, the Global System for Mobile Communications- Railway (GSM-R) is widely used in high-speed railway, but it is a 2G narrowband system that cannot meet the needs of intelligent development of high-speed railways. In the future, space-ground integrated railway communication network will gradually become an inevitable trend of railway development. Aiming at the problems of identity non-mutual

Traditional methods for analyzing Broken Access Control (BAC) vulnerabilities have limitations regarding low coverage of access control rules, high false positive rate (FPR). Additionally, state-of-the-art strategies for repairing BAC vulnerabilities utilizing statement-level replacement as a repair method may introduce new logical errors. To address these challenges, we propose a novel approach called

The management of complex, dynamic, and cross-domain resources in cyber-physical-human systems (CPHS) faces significant challenges under spatiotemporal dynamics, particularly resource state conflicts caused by rapid environmental changes and interdependent resource interactions. To address these challenges, this study proposes an integrated framework combining resource modeling and resource state adaptive

The large action space is a key challenge in reinforcement learning. Although hierarchical methods have been proven to be effective in addressing this issue, they are not fully explored. This paper combines domain knowledge with hierarchical concepts to propose a novel Hierarchical Reinforcement Learning framework based on macro actions (HRL-MA). This framework includes a macro action mapping model

Fleet formation identification in remote sensing is a significant focus in maritime surveillance. However, fleet may occur with different ship dense and noisy data due to the complex background and different satellite resolution, few studies have discussed formation identification considering the limits of sensing and application. This study introduces an effective fleet formation identification and

Cross-resolution Person Re-identification (ReID) faces the significant challenge of large resolution variance across different camera views in real surveillance systems. Most approaches based on super-resolution (SR) excessively rely on the SR images, which may lead to the loss of low-resolution (LR) information. Meanwhile, the region-agnostic SR could pose interference to ReID. For this, we propose

Open Set Domain Adaptation (OSDA) aims to minimize domain variation while distinguishing between known and unknown samples. However, existing OSDA methods, which rely on deep neural network classifiers, often lead to overconfident predictions and fail to clearly demarcate known from unknown samples. To address this limitation, we propose the Energy-based Open Set domain adaptation (EOS) method. EOS

Differential evolution (DE) is widely recognized as one of the most potent optimization algorithms, capable of effectively addressing a broad spectrum of optimization challenges. Nevertheless, even the most advanced variants of DE share some common challenges. This paper introduces a novel multi-stage semi-adaptive DE algorithm with enhanced diversity (MSA-DE), offering several key contributions: first

Detecting network anomalies is a critical cybersecurity task, yet existing methods struggle with high-dimensional data and limited interpretability in latent space. These challenges hinder precise differentiation between normal and anomalous activities due to (i) the chaotic distribution of normal samples, (ii) the absence of constraints to optimize the normal region’s hypervolume leading to high false

In current blockchain verifiable query research, redundant storage of data to be indexed is often required to enable efficient and feature-rich query algorithms. However, most blockchains currently face the problem of rapid data growth, leading to significant storage resource consumption by nodes. To provide a high-efficiency and generic verifiable query capability while reducing the storage burden

Federated Learning (FL) aims to allow numerous clients to participate in collaborative training in an efficient communication manner without exchanging private data. Traditional FL assumes that all clients have sufficient local resources to train models with the same architecture, and does not consider the reality that clients may struggle to deploy the same model across devices with varying computational

Classifying vibration mode shapes of a structure in an engineering design cycle can be a labor intensive and repetitive task. Although several methods have been proposed to automatically classify mode shapes, most existing models cannot fully represent mode shapes using both structural and modal information, limiting their application to specific structures. In this paper, we propose a graph convolutional