Search

With the increasing awareness of privacy protection and the improvement of relevant laws, federal learning has gradually become a new choice for cross-agency and cross-device machine learning. In order to solve the problems of privacy leakage, high computational overhead and high traffic in some federated learning schemes, this paper proposes a multiplicative double privacy mask algorithm which is convenient for homomorphic addition aggregation. The combination of homomorphic encryption and secret sharing ensures that the server cannot compromise user privacy from the private gradient uploaded by the participants. At the same time, the proposed TQRR (Top-Q-Random-R) gradient selection algorithm is used to filter the gradient of encryption and upload efficiently, which reduces the computing overhead of 51.78% and the traffic of 64.87% on the premise of ensuring the accuracy of the model, which makes the framework of privacy protection federated learning lighter to adapt to more miniaturized federated learning terminals. [ABSTRACT FROM AUTHOR]

Background: Blood biomarkers are crucial for the diagnosis and therapy of Alzheimer's disease (AD). Energy metabolism disturbances are closely related to AD. However, research on blood biomarkers related to energy metabolism is still insufficient. Objective: This study aims to explore the diagnostic and therapeutic significance of energy metabolism-related genes in AD. Methods: AD cohorts were obtained from GEO database and single center. Machine learning algorithms were used to identify key genes. GSEA was used for functional analysis. Six algorithms were utilized to establish and evaluate diagnostic models. Key gene-related drugs were screened through network pharmacology. Results: We identified 4 energy metabolism genes, NDUFA1, MECOM, RPL26, and RPS27. These genes have been confirmed to be closely related to multiple energy metabolic pathways and different types of T cell immune infiltration. Additionally, the transcription factors INSM2 and 4 lncRNAs were involved in regulating 4 genes. Further analysis showed that all biomarkers were downregulated in the AD cohorts and not affected by aging and gender. More importantly, we constructed a diagnostic prediction model of 4 biomarkers, which has been validated by various algorithms for its diagnostic performance. Furthermore, we found that valproic acid mainly interacted with these biomarkers through hydrogen bonding, salt bonding, and hydrophobic interaction. Conclusions: We constructed a predictive model based on 4 energy metabolism genes, which may be helpful for the diagnosis of AD. The 4 validated genes could serve as promising blood biomarkers for AD. Their interaction with valproic acid may play a crucial role in the therapy of AD. [ABSTRACT FROM AUTHOR]

The digitalization of production in smart grids entails challenges related to data collection, coordination, privacy protection, and anomaly detection. Machine learning techniques offer effective tools for processing Big Data, but identifying critical system states amidst vast amounts of data remains a challenge. To expedite data analysis, preprocessing through machine learning algorithms becomes essential. This paper introduces the advanced FedSVD algorithm, utilizing Singular Value Decomposition (SVD), which efficiently decomposes large datasets, establishes relationships, and identifies irrelevant data. The algorithm operates in federated machine learning systems, enabling local data processing on private devices while sharing only results with the global learning model. This approach enhances information processing confidentiality and facilitates the exchange of anomaly detection outcomes among network devices. The results of the study demonstrate that the modified FedSVD processing is 5 ms faster on average in comparison to the non-modified one. The proposed FedSVD algorithm calculates anomaly detection with higher accuracy by an average of 1–3% compared to the non-modified FedSVD and SVD ones. The advanced FedSVD algorithm proves to be a decentralized, confidential, and efficient solution for anomaly detection in smart grid systems.

BACKGROUND: Coronary heart disease (CHD) is the first cause of death globally. Hypertension is considered to be the most important independent risk factor for CHD. Early and accurate diagnosis of CHD in patients with hypertension can plays a significant role in reducing the risk and harm of hypertension combined with CHD. OBJECTIVE: To propose a non-invasive method for early diagnosis of coronary heart disease according to tongue image features with the help of machine learning techniques. METHODS: We collected standard tongue images and extract features by Diagnosis Analysis System (TDAS) and ResNet-50. On the basis of these tongue features, a common machine learning method is used to customize the non-invasive CHD diagnosis algorithm based on tongue image. RESULTS: Based on feature fusion, our algorithm has good performance. The results showed that the XGBoost model with fused features had the best performance with accuracy of 0.869, the AUC of 0.957, the AUPR of 0.961, the precision of 0.926, the recall of 0.806, and the F1-score of 0.862. CONCLUSION: We provide a feasible, convenient, and non-invasive method for the diagnosis and large-scale screening of CHD. Tongue image information is a possible effective marker for the diagnosis of CHD. [ABSTRACT FROM AUTHOR]

Background and Aims: A reduction in hepatic venous pressure gradient (HVPG) is the most accurate marker for assessing the severity of portal hypertension and the effectiveness of intervention treatments. This study aimed to evaluate the prognostic potential of blood‐based proteomic biomarkers in predicting HVPG response amongst cirrhotic patients with portal hypertension due to Hepatitis C virus (HCV) and had achieved sustained virologic response (SVR). Methods: The study comprised 59 patients from two cohorts. Patients underwent paired HVPG (pretreatment and after SVR), liver stiffness (LSM), and enhanced liver fibrosis scores (ELF) measurements, as well as proteomics‐based profiling on serum samples using SomaScan® at baseline (BL) and after SVR (EOS). Machine learning with feature selection (Caret, Random Forest and RPART) methods were performed to determine the proteins capable of classifying HVPG responders. Model performance was evaluated using AUROC (pROC R package). Results: Patients were stratified by a change in HVPG (EOS vs. BL) into responders (greater than 20% decline in HVPG from BL, or <10 mmHg at EOS with >10 mmHg at BL) and non‐responders. LSM and ELF decreased markedly after SVR but did not correlate with HVPG response. SomaScan (SomaLogic, Inc., Boulder, CO) analysis revealed a substantial shift in the peripheral proteome composition, reflected by 82 significantly differentially abundant proteins. Twelve proteins accurately distinguished responders from non‐responders, with an AUROC of.86, sensitivity of 83%, specificity of 83%, accuracy of 83%, PPV of 83%, and NPV of 83%. Conclusions: A combined non‐invasive soluble protein signature was identified, capable of accurately predicting HVPG response in HCV liver cirrhosis patients after achieving SVR. [ABSTRACT FROM AUTHOR]

Turbofan engines are known as the heart of the aircraft. The turbofan's health state determines the aircraft's operational status. Therefore, the equipment monitoring and maintenance of the engine is an important part of ensuring the healthy and stable operation of the aircraft, and it is vital to monitor the remaining useful life (RUL) of the engine. The monitored data of turbofan engines have high dimensions and a long time span, which cause difficulties in predicting the remaining useful life of the engine. This paper proposes a residual life prediction model based on Autoencoder and a Temporal Convolutional Network (TCN). Among them, Autoencoder is used to reduce the dimension of the data and extract features from the engine monitoring data. The TCN network is trained on the obtained low-dimensional data to predict the remaining useful life. The model mentioned in this article is verified on the NASA public data set (C-MAPSS) and compared with common machine learning methods and other deep neural networks. The SAE-TCN model achieved better scores on the FD001 independent testing data set with an RMSE of 18.01 and a score of 161. The average relative error of the model relative to other common learning models is 0.9499 in RMSE and 0.2656 in Scoring Function. The experimental results show that the model proposed in this paper performs the best in the evaluation, and this conclusion has important implications for engine health. [ABSTRACT FROM AUTHOR]

Web API is an efficient way for Service-based Software (SBS) development, and mashup is a key technology which merges several web services to deal with the increasing complexity of software requirements and expedite the service-based system development. The efficient service recommendation method is vital for the software development. However, the existing methods often suffer from data sparsity or cold start issues, which should lead to bad effects. Currently, this paper starts with SBS development, and proposes a service recommendation method based on knowledge graph embedding and collaborative filtering (CF) technology. In our model, we first construct a refined knowledge graph using SBS-service co-invocation record and SBS and service related information to mine the potential semantics relationship between SBS and service. Then, we learn the SBS and service entities in the knowledge graph. These heterogeneous entities (SBS and service, etc.) are embedded into the low-dimensional space through the representation learning algorithms of Word2vec and TransR, and the distances between SBS and service vectors are calculated. The input of recommendation model is SBS requirement (target SBS), the similarities functional SBS set is extracted from knowledge graph, which can relieve the cold start problem. Meanwhile, the recommendation model uses CF to recommend service to target SBS. Finally, this paper verifies the effectiveness of method on the real-word dataset. Compared with the several state-of-the-art methods, our method has the best service hit rate and ranking quality. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI) has been steadily developing in the medical field in the past few years, and AI-based applications have advanced cancer diagnosis. Breast cancer has a massive amount of data in oncology. There has been a high level of research enthusiasm to apply AI techniques to assist in breast cancer diagnosis and improve doctors' efficiency. However, the wise utilization of tedious breast cancer-related medical care is still challenging. Over the past few years, AI-based NLP applications have been increasingly proposed in breast cancer. In this systematic review, we conduct the review using preferred reporting items for systematic reviews and meta-analyses (PRISMA) and investigate the recent five years of literature in natural language processing (NLP)-based AI applications. This systematic review aims to uncover the recent trends in this area, close the research gap, and help doctors better understand the NLP application pipeline. We first conduct an initial literature search of 202 publications from Scopus, Web of Science, PubMed, Google Scholar, and the Association for Computational Linguistics (ACL) Anthology. Then, we screen the literature based on inclusion and exclusion criteria. Next, we categorize and analyze the advantages and disadvantages of the different machine learning models. We also discuss the current challenges, such as the lack of a public dataset. Furthermore, we suggest some promising future directions, including semi-supervised learning, active learning, and transfer learning. [ABSTRACT FROM AUTHOR]

Due to the similar chemical composition and matrix effect, the accurate identification of mineral pigments on wall paintings has brought great challenges. This work implemented an identification study on three mineral pigments with similar chemical compositions by combining LIBS technology with the K-nearest neighbor algorithm (KNN), random forest (RF support vector machine (SVM), back propagation artificial neural network (Bp-ANN) and convolutional neural network (CNN) to find the most suitable identification method for mural research. Using the SelectKBest algorithm, 300 characteristic lines with the largest difference among the three pigments were determined. The identification models of KNN, RF, SVM, Bp-ANN and CNN were established and optimized. The results showed that, except for the KNN model, the identification accuracy of other models for mock-up mural samples was above 99%. However, only the identification accuracy of 2D-CNN models reached above 94% for actual mural samples. Therefore, the 2D-CNN model was determined as the most suitable model for the identification and analysis of mural pigments. [ABSTRACT FROM AUTHOR]

The Chimp Optimization Algorithm (ChOA) is a heuristic algorithm proposed in recent years. It models the cooperative hunting behaviour of chimpanzee populations in nature and can be used to solve numerical as well as practical engineering optimization problems. ChOA has the problems of slow convergence speed and easily falling into local optimum. In order to solve these problems, this paper proposes a novel chimp optimization algorithm with refraction learning (RL-ChOA). In RL-ChOA, the Tent chaotic map is used to initialize the population, which improves the population's diversity and accelerates the algorithm's convergence speed. Further, a refraction learning strategy based on the physical principle of light refraction is introduced in ChOA, which is essentially an Opposition-Based Learning, helping the population to jump out of the local optimum. Using 23 widely used benchmark test functions and two engineering design optimization problems proved that RL-ChOA has good optimization performance, fast convergence speed, and satisfactory engineering application optimization performance. [ABSTRACT FROM AUTHOR]

Exploring lead‐free candidates and improving efficiency and stability remain the obstacle of hybrid organic‐inorganic perovskite‐based devices commercialization. Traditional trial‐and‐error methods seriously restrict the discovery especially for large search space, complex crystal structure and multi‐objective properties. Here, the authors propose a multi‐step and multi‐stage screening scheme to accelerate the discovery of hybrid organic‐inorganic perovskites A2BB′X6 from a large number of candidates through combining machine learning with high‐throughput calculations for pursuing excellent efficiency and thermal stability in solar cells. Followed by a series of screenings, the structure‐property relationships mapping A2BB′X6 properties are built and the predictions are close to reported experimental results. Successfully, four experimental‐feasibly candidates with good stability, high Debye temperature and suitable band gap are screened out and further verified by density‐functional theory calculations, in which the predicted efficiency for three lead‐free candidates ((CH3NH3)2AgGaBr6, (CH3NH3)2AgInBr6 and (C2NH6)2AgInBr6) achieves 20.6%, 19.9% and 27.6% due to ultrabroadband absorption region ranging from UVC to IRC with excitonic radiative combination rates as low as 10 ps, large or intermediate polarons form with properties similar to CH3NH3PbI3 and the calculated thermal conductivities are 5.04, 4.39 and 5.16 Wm−1K−1, respectively, with Debye temperatures larger than 500 K, beneficial for suppression of both nonradiative combination and heat‐induced degradation. [ABSTRACT FROM AUTHOR]

Additive manufacturing usually involves the complex interaction of design, materials, and manufacturing, often resulting in long and cost-intensive iterative evaluation cycles. Therefore, it is critical for it to be aligned with the Materials Genome Initiative to develop, produce, and deploy high-throughput components. Recognizing a need, this study leverages a feature-based qualification (FBQ) methodology to decompose a complex structure by identifying critical performance-limiting features, for the purpose of reducing the cost and time of DED process qualification. A hybrid-physics-based multi-objective optimization tool was used to predict processing-structure-property relationships in thin-walled builds. The probabilistic ML models achieved targeted predictions with half the sample space when compared with conventional DOEs, while also being 37–50% more reliable with respect to regression tools with linear basis function. Although the current model developed is specific to Ti64 builds in a RPMi557 powder-feed DED machine, the FBQ methodology may be more universally employed to other material-modality combinations. [ABSTRACT FROM AUTHOR]

Transparent intumescent fire retardant coatings (IFRC) are widely used to protect ancient buildings and high-end furniture from fire. Currently, IFRC fraud has attracted considerable attention due to the lack of efficient techniques to detect the identity of IFRC products. In this paper, we use Raman spectroscopy and machine learning to identify IFRC brands. A total of 135 transparent IFRC samples of 6 common brands were prepared and scanned using a portable Raman spectrometer. The obtained spectra were pre-processed to remove unwanted variation and classified using partial least squares discriminant analysis (PLS-DA) and kernel extreme learning machine (K-ELM). The accuracies achieved by PLS-DA and K-ELM were 0.978 and 0.993, respectively. In addition, important variables were determined based on the regression coefficients of PLS-DA. The results demonstrate that portable Raman spectroscopy combined with machine learning is a viable approach for rapid, on-situ and low-cost identification of IFRC brands. • Portable Raman spectroscopy for brand identification of intumescent fire retardant coating. • Classification accuracy of 0.978–0.993 achieved by machine learning methods. • A fast, on-situ and low-cost approach for detecting intumescent fire retardant coating. [ABSTRACT FROM AUTHOR]

The World Health Organization (WHO) declared on 11th March 2020 the spread of the coronavirus disease 2019 (COVID-19) a pandemic. The traditional infectious disease surveillance had failed to alert public health authorities to intervene in time and mitigate and control the COVID-19 before it became a pandemic. Compared with traditional public health surveillance, harnessing the rich data from social media, including Twitter, has been considered a useful tool and can overcome the limitations of the traditional surveillance system. This paper proposes an intelligent COVID-19 early warning system using Twitter data with novel machine learning methods. We use the natural language processing (NLP) pre-training technique, i.e., fine-tuning BERT as a Twitter classification method. Moreover, we implement a COVID-19 forecasting model through a Twitter-based linear regression model to detect early signs of the COVID-19 outbreak. Furthermore, we develop an expert system, an early warning web application based on the proposed methods. The experimental results suggest that it is feasible to use Twitter data to provide COVID-19 surveillance and prediction in the US to support health departments' decision-making. [ABSTRACT FROM AUTHOR]

In recent years, artificial intelligence (AI) has shown great promise in medicine. However, explainability issues make AI applications in clinical usages difficult. Some research has been conducted into explainable artificial intelligence (XAI) to overcome the limitation of the black-box nature of AI methods. Compared with AI techniques such as deep learning, XAI can provide both decision-making and explanations of the model. In this review, we conducted a survey of the recent trends in medical diagnosis and surgical applications using XAI. We have searched articles published between 2019 and 2021 from PubMed, IEEE Xplore, Association for Computing Machinery, and Google Scholar. We included articles which met the selection criteria in the review and then extracted and analyzed relevant information from the studies. Additionally, we provide an experimental showcase on breast cancer diagnosis, and illustrate how XAI can be applied in medical XAI applications. Finally, we summarize the XAI methods utilized in the medical XAI applications, the challenges that the researchers have met, and discuss the future research directions. The survey result indicates that medical XAI is a promising research direction, and this study aims to serve as a reference to medical experts and AI scientists when designing medical XAI applications. [ABSTRACT FROM AUTHOR]

Alzheimer's Disease (AD) is a neurodegenerative disease that commonly occurs in older people. It is characterized by both cognitive and functional impairment. However, as AD has an unclear pathological cause, it can be hard to diagnose with confidence. This is even more so in the early stage of Mild Cognitive Impairment (MCI). This paper proposes a U-Net based Generative Adversarial Network (GAN) to synthesize fluorodeoxyglucose - positron emission tomography (FDG-PET) from magnetic resonance imaging - T1 weighted imaging (MRI-T1WI) for further usage in AD diagnosis including its early-stage MCI. The experiments have displayed promising results with Structural Similarity Index Measure (SSIM) reaching 0.9714. Furthermore, three types of classifiers are developed, i.e., one Multi-Layer Perceptron (MLP) based classifier, two Graph Neural Network (GNN) based classifiers where one is for graph classification and the other is for node classification. 10-fold cross-validation has been conducted on all trials of experiments for classifier comparison. The performance of these three types of classifiers has been compared with the different input modalities setting and data fusion strategies. The results have shown that GNN based node classifier surpasses the other two types of classifiers, and has achieved the state-of-the-art (SOTA) performance with the best accuracy at 90.18% for 3-class classification, namely AD, MCI and normal control (NC) with the synthesized fluorodeoxyglucose - positron emission tomography (FDG-PET) features fused at the input level. Moreover, involving synthesized FDG-PET as part of the input with proper data fusion strategies has also proved to enhance all three types of classifiers' performance. This work provides support for the notion that machine learning-derived image analysis may be a useful approach to improving the diagnosis of AD. [ABSTRACT FROM AUTHOR]

Image captioning enables people to better understand images through fine-grained analysis. Recently the encoder-decoder architecture with attention mechanism has achieved great achievements in image captioning and visual question answering. In this paper, we propose a new captioning algorithm that integrates two separate LSTM (Long-short Term Memory) networks through an adaptive semantic attention model. Within our approach, the first LSTM network is followed by an attention model, which serves as a visual sentinel can flexibly make a trade off between the visual semantic region and textual content. Another LSTM is used as a language model, which combines the hidden state representation of the first LSTM and attention context vector, then outputs the word sequence. The proposed model has been extensively evaluated on two large-scale datasets: MSCOCO and Flickr30k. Experimental results show that the proposed method pays more attention to visual salient regions and achieves significant performance of prior state-of-the-art approaches on multiple evaluation metrics. [ABSTRACT FROM AUTHOR]

Efficient determination of the rock mass properties is vitally important for calculating and evaluating tunnel stability in tunnel engineering. The back analysis method has been widely used as an indirect method for determining rock mass parameters based on field measurements. However, most back analysis methods are generally time-consuming for numerical simulation and are merely based on the measured displacement, which leads to the identification of rock mass parameters that cannot fully reflect the characteristics of the surrounding rock. To improve the accuracy of the estimation of rock mass parameters, this paper presents a back analysis method based on multi-output support vector regression (MSVR) and differential evolution (DE) algorithms. Firstly, the global sensitivity analysis of rock mass parameters is analyzed using the elementary effects method. Numerical simulation is then carried out to prepare training samples. DE algorithm is used to determine the optimum hyperparameters of MSVR. Based on the monitoring data, the rock mass properties of the selected sensitive parameters are estimated by the constructed MSVR model. A high-speed railway tunnel is utilized to demonstrate the effectiveness of the MSVR with the DE algorithm (DE-MSVR). The results show that the DE-MSVR with mixed monitoring data of vault settlement, convergence, and rock mass stress has higher forecasting performance than these models with a single type of monitoring data. It is feasible to use the monitoring data at the early stages combined with the numerical simulation for parameter back analysis. Moreover, the comparison results show that the presented method exhibits higher prediction accuracy than the existing back analysis models. [ABSTRACT FROM AUTHOR]

Due to its easy implementation and comprehensive applicability, surrogate model has been widely used in the aerodynamic design optimization (ADO) of turbomachinery blades. Machine learning method, rapidly developed and applied in many disciplines, is beneficial to model training. The paper works on ADO of a transonic fan rotor by blade sweeping using a supervised learning model method. First, verifications of numerical simulation and specifications of blade design optimization are given. Then the principles and implementations of an adaptive Gaussian Process (GP) are introduced. The acquisition function for adaptive sampling firstly increases and then decreases and ultimately converges in the process of model update. Meanwhile, performance predictions of the swept blades are achieved using GP model. From probability estimation and Sobol sensitivity analysis, some priori design knowledge revealing the impacts of blade sweeping on aerodynamic performance is preliminarily discovered. The results demonstrate that the concerned performance parameters are sensitive to the design parameters on the outer spans of the rotor blade. Finally, multi-objective optimization maximizing the total pressure ratio and adiabatic efficiency with constrained mass flow rate is carried out and the Pareto front is determined. The results of the original and three swept blades are compared in detail to illustrate the effects of blade sweeping to flow variations. For the transonic rotor, forward sweeping favors the increase of total pressure ratio, while backward sweeping on the outer spans favors the increase of adiabatic efficiency. [ABSTRACT FROM AUTHOR]

Shear strength and ductile deformation capacity of monolithic exterior shear keys play important roles on the seismic behavior of highway bridges. Improving the ductility of the shear key through reasonable design could effectively restrict superstructure displacement and minimize damage to pier columns. However, most existing studies primarily focus on evaluating the shear strength and damage mechanisms of shear keys, with limited attention given to analytical models for assessing ductile deformation capacity. To this end, this study proposes a method that combines quantile regression and machine learning techniques to predict the ductile deformation limit states of shear keys. The proposed numerical modeling method for shear keys is validated using the experimental data on lateral load-displacement curves. On this basis, Latin hypercubic sampling and joint simulation are utilized to establish the ductile deformation limit state database. The ductile deformation limit states of shear keys are characterized by their lateral displacements in three distinct damage states. The main influencing factors of each limit states are investigated by correlation analysis. Three quantile regression machine learning models are trained and constructed to conduct point and interval prediction for the ductile deformation limit states. All three models, particularly the Gaussian process quantile regression model, are proven to demonstrate exceptional prediction accuracy and uncertainty quantification. Experimental data validation confirms that these models outperform the prediction accuracy of existing simplified models. • Finite element modeling approach of shear keys is developed and validated. • Main factors of ductile limit states are investigated by correlation analysis. • The QR-ML techniques are adopted to establish probabilistic prediction models. • The established models are compared for point and interval prediction performance. [ABSTRACT FROM AUTHOR]

[Display omitted] • The dominant features affecting the band gap are screened via machine learning. • A mapping strategy for band gap range classification is developed. • A guiding principle is proposed for screening potential halide perovskites with suitable band gap ranges. Recently, perovskite solar cells (PSCs) have received great attentions as the most promising candidate for the next generation solar cells; where the halide perovskite with ABX 3 stoichiometry are playing as key components. As a key relevant parameter for various applications of perovskites, the band gap can be modified and optimized by tuning their compositions. In order to enhance the screening efficiency of perovskites with appropriate band gap range for particular applications, this work developed a mapping strategy for band gap range classification through data-driven selected features. This work demonstrates that the proposed features, and further the developed band gap maps, are able to offer a useful initial guiding principle for screenings of potential halide perovskites with fitting band gap ranges and provide opportunities for their compositional design. [ABSTRACT FROM AUTHOR]

This study develops a novel efficient coupled distance metric learning algorithm called coupled marginal discriminant mappings (CMDM). It can provide a low-resolution face recognition method with coupled mappings, making the data points in the original high and low resolution features projected into a unified space, where classification is implemented. At the same time, it makes samples from the same class gather more closely while makes samples of distinct class disperse more separately with a large margin. Thus, for the low-resolution face recognition issue, the proposed method can be leveraged to avoid dimensional mismatch problem and fill different-resolution data gap. The experimental evaluation based on the AR and FERET face databases demonstrates that CMDM can achieve highly competitive performance comparing favorably with the existing state-of-the-art low-resolution face recognition methods. [ABSTRACT FROM AUTHOR]