Accurate kinetic modeling of hydrocracking, a process that converts heavy hydrocarbons into lighter, high-value products, is crucial for refinery optimization and energy efficiency. Traditional lumped mechanistic models often suffer from parameter identifiability issues, while purely data-driven methods ignore physical laws, limiting interpretability and extrapolation. This work presents a Physics Informed Neural Ordinary Differential Equation (PINODE) framework that integrates reaction stoichiometry, rate positivity, and Arrhenius temperature dependence directly into the learning architecture. Using synthetic datasets from two established five-lump hydrocracking models, multiple PINODE variants were benchmarked, demonstrating that progressively embedding physical constraints reduces concentration prediction errors by 40%–55% relative to an unconstrained baseline. The best-performing variants achieved test RMSE (Root Mean Squared Error) values below 0.25 and recovered activation energies within acceptable limits of literature values. Validation against semi-empirical and experimental data from Coronel-García et al. confirmed that the framework generalizes beyond synthetic conditions, reproducing measured product trends and temperature effects. Bootstrap resampling and 95% confidence intervals from Arrhenius regressions established the statistical robustness of the recovered parameters under noise and data sparsity. Overall, the PINODE framework serves as an interpretable and statistically reliable grey-box modeling tool, bridging mechanistic and data-driven approaches for kinetic modeling and refinery process optimization. • Physics-Informed Neural ODEs with embedded physics like stoichiometry and Arrhenius laws. • Demonstrates strong predictive accuracy and 40%–55% RMSE reduction vs. unconstrained NODE. • Validated against experimental data via a semi-empirical surrogate with bootstrap confidence intervals. • Recovers physically meaningful kinetic parameters with 95% uncertainty bounds.

Recently, Moore-Penrose inverse (MPI)-based parameter fine-tuning of fully connected (FC) layers in pretrained deep convolutional neural networks (DCNNs) has emerged within the inductive transfer learning (ITL) paradigm. However, this approach has not gained significant traction in practical applications due to its stringent computational requirements. This work addresses this issue through a novel fast retraining strategy that enhances applicability of the MPI-based ITL. Specifically, during each retraining epoch, a random layer freezing protocol is utilized to manage the number of layers undergoing feature refinement. Additionally, this work incorporates an MPI-based approach for refining the trainable parameters of FC layers under batch processing, contributing to expedited convergence. Extensive experiments on several ImageNet pretrained benchmark DCNNs demonstrate that the proposed ITL achieves competitive performance with excellent convergence speed compared to conventional ITL methods. For instance, the proposed strategy converges nearly 1.5 times faster than retraining the ImageNet pretrained ResNet-50 using stochastic gradient descent with momentum (SGDM).

The broad learning system (BLS) is a versatile and effective tool for analyzing tabular data. However, the rapid expansion of big data has resulted in an overwhelming amount of tabular data, necessitating the development of specialized tools for effective management and analysis. This article introduces an optimized BLS (OBLS) specifically tailored for big data analysis. In addition, a deep-optimized BLS (DOBLS) network is developed further to enhance the performance and efficiency of the OBLS. The main contributions of this article are: 1) by retracing the network's error from the output space to the latent space, the OBLS adjusts parameters in the feature and enhancement node layers. This process aims to achieve more resilient representations, resulting in improved performance; 2) the DOBLS is a multilayered structure consisting of multiple OBLSs, wherein each OBLS connects to the input and output layers, enabling direct data propagation. This design helps reduce information loss between layers, ensuring an efficient flow of information throughout the network; and 3) the proposed methods demonstrate robustness across various applications, including multiview feature embedding, one-class classification (OCC), camera model identification, electroencephalogram (EEG) signal processing, and radar signal analysis. Experimental results validate the effectiveness of the proposed models. To ensure reproducibility, the source code is available at https://github.com/1027051515/OBLS_DOBLS.

High-frequency surface wave radar (HFSWR) is a powerful tool for ship detection and surveillance. blackHowever, the use of pre-trained deep learning (DL) networks for ship detection is challenging due to the limited training samples in HFSWR and the substantial differences between remote sensing images and everyday images. To tackle these issues, this paper proposes a coarse-to-fine target detection approach that combines traditional methods with DL, resulting in improved performance. The contributions of this work include: 1) a two-stage learning pipeline that integrates spatial-frequency analysis (SFA) with subnet-based neural networks, 2) an automatic linear thresholding algorithm for plausible target region (PTR) detection, and 3) a robust subnet neural network for fine target detection. The advantage of using SFA and subnet network is that the SFA reduces the need for extensive training data, while the subnet neural network excels at localizing ships even with limited training data. Experimental results on the HFSWR-RD dataset affirm the model's superior performance compared to rival algorithms.

A high-frequency surface wave radar (HFSWR) is an effective tool for monitoring an exclusive economic zone (EEZ). However, the presence of diverse clutters and noises that contaminate the echo signals of the radar hinder its maritime surveillance. To address this issue, this paper presents a two-stage hierarchical one-class classification network (HOCN) designed specifically for ship-target detection in range-Doppler (RD) images. In Stage 1, the plausible region of interest (PROI) is extracted. This stage employs a dynamic threshold optimization strategy and Laplacian kernel to identify the potential regions of interest. In Stage 2, the proposed one-class deconvolutional-and-convolutional network (OC-DCNet) is utilized for fine detection of ship-targets. This stage comprises two sub-modules: the deconvolutional sub-module, which expands the input into a 2D matrix, and the convolutional sub-module, which classifies the input pattern as either a ship-target or a non-ship-target. The experimental results on a newly collected dataset called HFRD demonstrate the effectiveness of the proposed HFSWR ship-target detection algorithm.

Summary Due to the development of mobile transmission techniques, more exciting services can be launched that may change our daily lives in more easy and comfortable ways. The introduction of device‐to‐device (D2D) service in future 5G communication systems can extend the service coverage of intelligent medical treatment from hospitals to anywhere inside the network coverage area, which can save precious time and medical resources. As one of the key techniques in a 5G system, D2D technology has attracted extensive attention from academia and industry due to its system performance improvement, user experience enhancement and service extension. However, the introduction of D2D service will tremendously pollute the system transmission environment due to the induced interference, especially in the downlink direction where the interferer is very strong. This situation brings new challenges for effective and green resource utilization. To solve this problem, a green spectrum resource allocation strategy based on the Hungarian method is proposed in this paper to optimize system spectrum efficiency while considering the fairness among users under the assumption that all resources are fully shared by traditional cellular and newly introduced D2D users. To validate the performance of the proposed algorithm, a system‐level Monte Carlo simulation is also conducted, which shows the favorable performance of the proposed algorithm over the traditional greedy algorithm.

Web attacks such as Cross-Site Scripting and SQL Injection are serious Web threats that lead to catastrophic data leaking and loss. Because attack payloads are often short segments hidden in URL requests/posts that can be very long, classical machine learning approaches have difficulties in learning useful patterns from them. In this study, we propose a novel Locate-Then-Detect (LTD) system that can precisely detect Web threats in real-time by using attention-based deep neural networks. Firstly, an efficient Payload Locating Network (PLN) is employed to propose most suspicious regions from large URL requests/posts. Then a Payload Classification Network (PCN) is adopted to accurately classify malicious regions from suspicious candidates. In this way, PCN can focus more on learning malicious segments and highly increase detection accuracy. The noise induced by irrelevant background strings can be largely eliminated. Besides, LTD can greatly reduce computational costs (82.6% less) by ignoring large irrelevant URL content. Experiments are carried out on both benchmarks and real Web traffic. The LTD outperforms an HMM-based approach, the Libinjection system, and a leading commercial rule-based Web Application Firewall. Our method can be efficiently implemented on GPUs with an average detection time of about 5ms and well qualified for real-time applications.