The development of data-driven soft sensors for modeling complex data, particularly in scenarios characterized by strong nonlinearity, high dimensionality, cross-correlation and autocorrelation, remains a significant ...
详细信息
ISBN:
(数字)9798350355642
ISBN:
(纸本)9798350355659
The development of data-driven soft sensors for modeling complex data, particularly in scenarios characterized by strong nonlinearity, high dimensionality, cross-correlation and autocorrelation, remains a significant challenge in process data analysis. Bidirectional Long Short-Term Memory (BiLSTM) networks are highly effective in capturing both forward and backward dependencies within sequences, thereby enhancing the extraction of temporal features from time series data. However, BiLSTM networks may struggle to capture global dependencies, especially in long sequences. To address this limitation, the self-attention mechanism, which excels in modeling global relationships across all positions in the input sequence, can be integrated with BiLSTM by applying weighted attention to the two different directions of BiLSTM, thereby enhancing its performance. This study proposes a novel method called SA-BiLSTM, which combines the strengths of the BiLSTM network and the self- attention mechanism. The self-attention layer captures global dependencies, while the BiLSTM layer extracts comprehensive temporal features in both directions, which are then passed to a fully connected layer for final prediction. Extensive comparisons with traditional methods such as BiLSTM, LSTM with added self-attention mechanism, LSTM, RNN, and SVM validate the superior performance of the proposed SA-BiLSTM method in handling complex time series data.
Poor atmospheric visibility is a leading cause of aviation accidents. This paper explores the "visibility supply chain";how these data are collected, how they are consumed, the current challenges, and effort...
详细信息
Different social networks are independent of each other, and identifying the same person in different social networks is of great significance for cross-network information dissemination and the construction of a comp...
详细信息
Machine Translation (MT) is a crucial sub-field of Artificial Intelligence (AI). Since its inception in the mid-20th century, it tackles the intricate challenge of translating text across languages. MT has seen remark...
详细信息
Machine Translation (MT) is a crucial sub-field of Artificial Intelligence (AI). Since its inception in the mid-20th century, it tackles the intricate challenge of translating text across languages. MT has seen remarkable progress, particularly over the past decade with the advent of deep learning (DL) techniques. Key DL methods include feed-forward deep neural networks (NN), convolutional NN (CNN), recurrent NN (RNN), long term/ short-term memory networks (LSTM), Gated recurrent units (GRU), attention mechanisms, Transformer model and autoencoders. These advancements have enabled the widespread use of MT in web-based translation services, mobile applications, and various MT platforms. Each addressing different challenges and excelling in various aspects of MT. This research offers a thorough review of DL techniques in MT, identifying the most utilized MT systems, their architectures, and performance outcomes. It also highlights the merits, and limitations of these methods. The findings reveal that DL techniques for MT using transform model is the prevailing paradigm. The research concludes with a discussion on potential future research directions.
This paper is concerned with the event-driven stabilization of Markov jump systems with disturbances based on disturbance observer (DO). First, a DO is employed to estimate the disturbance generated by an exogenous sy...
详细信息
This study is designed to develop Artificial Intelligence(AI)based analysis tool that could accurately detect COVID-19 lung infections based on portable chest x-rays(CXRs).The frontline physicians and radiologists suf...
详细信息
This study is designed to develop Artificial Intelligence(AI)based analysis tool that could accurately detect COVID-19 lung infections based on portable chest x-rays(CXRs).The frontline physicians and radiologists suffer from grand challenges for COVID-19 pandemic due to the suboptimal image quality and the large volume of *** this study,AI-based analysis tools were developed that can precisely classify COVID-19 lung *** available datasets of COVID-19(N=1525),non-COVID-19 normal(N=1525),viral pneumonia(N=1342)and bacterial pneumonia(N=2521)from the Italian Society of Medical and Interventional Radiology(SIRM),Radiopaedia,The Cancer Imaging Archive(TCIA)and Kaggle repositories were taken.A multi-approach utilizing deep learning ResNet101 with and without hyperparameters optimization was ***,the fea-tures extracted from the average pooling layer of ResNet101 were used as input to machine learning(ML)algorithms,which twice trained the learning *** ResNet101 with optimized parameters yielded improved performance to default *** extracted features from ResNet101 are fed to the k-nearest neighbor(KNN)and support vector machine(SVM)yielded the highest 3-class classification performance of 99.86%and 99.46%,*** results indicate that the proposed approach can be bet-ter utilized for improving the accuracy and diagnostic efficiency of *** proposed deep learning model has the potential to improve further the efficiency of the healthcare systems for proper diagnosis and prognosis of COVID-19 lung infection.
In this work, we develop a multi-hop packet delay bound violation model using Support vector machines (SVM) to predict the packet loss probability and end-to-end distortion for video streaming over multi-hop networks....
详细信息
In this work, we develop a multi-hop packet delay bound violation model using Support vector machines (SVM) to predict the packet loss probability and end-to-end distortion for video streaming over multi-hop networks. Based on this model, we formulate the resource allocation into a non-convex optimization problem which aims to minimize the overall video distortion while maintaining fairness between sessions. We solve this optimization problem using Lagrangian duality methods. Extensive experimental results demonstrate that, with this widely-used offline-training-online-estimation mechanism, the proposed model is potentially applicable to almost all network conditions and can provide fairly accurate estimation results as compared with other models with a given sample data set. The proposed optimization algorithm achieves more efficient resource allocation than existing schemes.
This paper proposes an interactive forest algorithm for planning single-tuned harmonic filters for electric power distribution systems. The planning of harmonic filters is a complex problem that must consider multiple...
详细信息
Sign language is a priceless means of communication for deaf and hard-of-hearing people to fully enable them to participate in society and interact with others. This study introduces a novel universal sign language sy...
详细信息
The zero padding (ZP) variants of orthogonal frequency-division multiplexing (OFDM) exhibit a lower bit error rate (BER) and higher energy efficiency compared to their cyclic prefix (CP) counterparts. However, the emp...
The zero padding (ZP) variants of orthogonal frequency-division multiplexing (OFDM) exhibit a lower bit error rate (BER) and higher energy efficiency compared to their cyclic prefix (CP) counterparts. However, the employment of ZP-OFDM demands strict time synchronization, which is challenging in the absence of pilots or CP. Moreover, time synchronization in OFDM systems is even more challenging when impulsive noise is present. It is well known that urban noise, which consists largely of impulsive noise generated by spark plugs used in internal combustion engines, switching and industrial activities, and discharge of high voltage distribution lines, has a strong influence on digital mobile communications. In this paper, we propose a new low-complexity approximate maximum likelihood (A-ML) timing offset (TO) estimator for ZP multiple-input multiple-output (MIMO)-OFDM in impulsive-noise environments. Performance comparison of the A-ML estimator with existing TO estimators demonstrates a superior performance in terms of lock-in probability with similar computational complexity. Also, compared to the optimal ML TO estimator, it offers a significantly lower computational complexity with negligible performance loss. The A-ML estimator can be employed for both frame and symbol synchronization.
暂无评论