Search Results (1,233)

Financial market forecasting remains challenging due to complex nonlinear dynamics and regime-dependent behaviors that traditional models struggle to capture effectively. This research introduces the Adaptive Financial Reservoir Network with Hypernetwork Flow (AFRN–HyperFlow) framework, a novel ensemble architecture integrating Echo State Networks, temporal convolutional networks, mixture density networks, adaptive Hypernetworks, and deep state-space models for enhanced financial time-series prediction. Through comprehensive feature engineering incorporating technical indicators, spectral decomposition, reservoir-based representations, and flow dynamics characteristics, the framework achieves superior forecasting performance across diverse market conditions. Experimental validation on 26,817 balanced samples demonstrates exceptional results with an F1-score of 0.8947, representing a 12.3% improvement over State-of-the-Art baseline methods, while maintaining robust performance across asset classes from equities to cryptocurrencies. The adaptive Hypernetwork mechanism enables real-time regime-change detection with 2.3 days average lag and 95% accuracy, while systematic SHAP analysis provides comprehensive interpretability essential for regulatory compliance. Ablation studies reveal Echo State Networks contribute 9.47% performance improvement, validating the architectural design. The AFRN–HyperFlow framework addresses critical limitations in uncertainty quantification, regime adaptability, and interpretability, offering promising directions for next-generation financial forecasting systems incorporating quantum computing and federated learning approaches. Full article

Accurate and robust spatial perception is fundamental for dynamic 3D city modeling and urban environmental sensing. High-resolution remote sensing data, particularly LiDAR point clouds, are pivotal for these tasks due to their lighting invariance and precise geometric information. However, processing and aligning sequential LiDAR point clouds in complex urban environments presents significant challenges: traditional point-based or feature-matching methods are often sensitive to urban dynamics (e.g., moving vehicles and pedestrians) and struggle to establish reliable correspondences. While deep learning offers solutions, current approaches for point cloud alignment exhibit key limitations: self-supervised losses often neglect inherent alignment uncertainties, and supervised methods require costly pixel-level correspondence annotations. To address these challenges, we propose UnMinkLO-Net, an end-to-end self-supervised LiDAR odometry framework. Our method is as follows: (1) we efficiently encode 3D point cloud structures using voxel-based sparse convolution, and (2) we model inherent alignment uncertainty via covariance matrices, enabling novel self-supervised loss based on uncertainty modeling. Extensive evaluations on the KITTI urban dataset demonstrate UnMinkLO-Net’s effectiveness in achieving highly accurate point cloud registration. Our self-supervised approach, eliminating the need for manual annotations, provides a powerful foundation for processing and analyzing LiDAR data within multi-sensor urban sensing frameworks. Full article

Global supply chains often face uncertainties in production lead times, fluctuating exchange rates, and varying tariff regulations, all of which can significantly impact total profit. To address these challenges, this study formulates a multi-country supply chain problem as a Semi-Markov Decision Process (SMDP), integrating both currency variability and tariff levels. Using a Q-learning-based method (SMART), we explore three scenarios: (1) wide currency gaps under a uniform tariff, (2) narrowed currency gaps encouraging more local sourcing, and (3) distinct tariff structures that highlight how varying duties can reshape global fulfillment decisions. Beyond these baselines we analyze uncertainty-extended variants and targeted sensitivities (quantity discounts, tariff escalation, and the joint influence of inventory holding costs and tariff costs). Simulation results, accompanied by policy heatmaps and performance metrics, illustrate how small or large shifts in exchange rates and tariffs can alter sourcing strategies, transportation modes, and inventory management. A Deep Q-Network (DQN) is also applied to validate the Q-learning policy, demonstrating alignment with a more advanced neural model for moderate-scale problems. These findings underscore the adaptability of reinforcement learning in guiding practitioners and policymakers, especially under rapidly changing trade environments where exchange rate volatility and incremental tariff changes demand robust, data-driven decision-making. Full article

This paper presents a Systematic Mapping Study (SMS) on data-driven approaches in flood forecasting from 2019 to 2024, a period marked by transformative developments in Deep Learning (DL) technologies. Analysing 363 selected studies, this paper provides an overview of the technological evolution in this field, methodological approaches, evaluation practices and geographical distribution of studies. The study revealed that meteorological and hydrological factors constitute approximately 76% of input variables, with rainfall/precipitation and water level measurements forming the core predictive basis. Long Short-Term Memory (LSTM) networks emerged as the dominant algorithm (21% of implementations), whilst hybrid and ensemble approaches showed the most dramatic growth (from 2% in 2019 to 10% in 2024). The study also revealed a threefold increase in publications during this period, with significant geographical concentration in East and Southeast Asia (56% of studies), particularly China (36%). Several research gaps were identified, including limited exploration of graph-based approaches for modelling spatial relationships, underutilisation of transfer learning for data-scarce regions, and insufficient uncertainty quantification. This SMS provides researchers and practitioners with actionable insights into current trends, methodological practices, and future directions in data-driven flood forecasting, thereby advancing this critical field for disaster management. Full article

Rehabilitation devices such as actuated lower-limb exoskeletons can provide essential mobility assistance for pediatric patients with gait impairments. Enhancing their control systems under conditions of user variability and dynamic disturbances remains a significant challenge, particularly in active-assist modes. This study presents a human-in-the-loop control architecture for a pediatric lower-limb exoskeleton, combining outer-loop admittance control with robust inner-loop trajectory tracking via a non-singular terminal sliding-mode (NSTSM) controller. Designed for active-assist gait rehabilitation in children aged 8–12 years, the exoskeleton dynamically responds to user interaction forces while ensuring finite-time convergence under system uncertainties. To enhance adaptability, we augment the inner-loop control with a twin delayed deep deterministic policy gradient (TD3) reinforcement learning framework. The actor–critic RL agent tunes NSTSM gains in real-time, enabling personalized model-free adaptation to subject-specific gait dynamics and external disturbances. The numerical simulations show improved trajectory tracking, with RMSE reductions of 27.82% (hip) and 5.43% (knee), and IAE improvements of 40.85% and 10.20%, respectively, over the baseline NSTSM controller. The proposed approach also reduced the peak interaction torques across all the joints, suggesting more compliant and comfortable assistance for users. While minor degradation is observed at the ankle joint, the TD3-NSTSM controller demonstrates improved responsiveness and stability, particularly in high-load joints. This research contributes to advancing pediatric gait rehabilitation using RL-enhanced control, offering improved mobility support and adaptive rehabilitation outcomes. Full article

High-voltage diodes, as key devices in power electronic systems, have important significance for system reliability and preventive maintenance in terms of storage life prediction. In this paper, we propose a hybrid modeling framework that integrates the Long Short-Term Memory Network (LSTM) and Transformer structure, and is hyper-parameter optimized by the Improved Artificial Bee Colony Algorithm (IABC), aiming to realize the high-precision modeling and prediction of high-voltage diode storage life. The framework combines the advantages of LSTM in time-dependent modeling with the global feature extraction capability of Transformer’s self-attention mechanism, and improves the feature learning effect under small-sample conditions through a deep fusion strategy. Meanwhile, the parameter type-aware IABC search mechanism is introduced to efficiently optimize the model hyperparameters. The experimental results show that, compared with the unoptimized model, the average mean square error (MSE) of the proposed model is reduced by 33.7% (from 0.00574 to 0.00402) and the coefficient of determination (R²) is improved by 3.6% (from 0.892 to 0.924) in 10-fold cross-validation. The average predicted lifetime of the sample was 39,403.3 h, and the mean relative uncertainty of prediction was 12.57%. This study provides an efficient tool for power electronics reliability engineering and has important applications for smart grid and new energy system health management. Full article

In complex dynamic environments, robotic arms face multiple challenges such as real-time environmental changes, high-dimensional state spaces, and strong uncertainties. Trajectory planning tasks based on deep reinforcement learning (DRL) suffer from difficulties in acquiring human expert strategies, low experience utilization (leading to slow convergence), and unreasonable reward function design. To address these issues, this paper designs a neural network-based expert-guided triple experience replay mechanism (NETM) and proposes an improved reward function adapted to dynamic environments. This replay mechanism integrates imitation learning’s fast data fitting with DRL’s self-optimization to expand limited expert demonstrations and algorithm-generated successes into optimized expert experiences. Experimental results show the expanded expert experience accelerates convergence: in dynamic scenarios, NETM boosts accuracy by over 30% and safe rate by 2.28% compared to baseline algorithms. Full article

This paper studies deep reinforcement learning (DRL)-based joint resource allocation and three-dimensional (3D) trajectory optimization for unmanned aerial vehicle (UAV)–ground access point (GAP) cooperative non-orthogonal multiple access (NOMA) communication in Industrial Internet of Things (IIoT) systems. Cooperative and non-cooperative users adopt different signal transmission strategies to meet diverse, task-oriented, quality-of-service requirements. Specifically, the DRL framework based on the Soft Actor–Critic algorithm is proposed to jointly optimize user scheduling, power allocation, and UAV trajectory in continuous action spaces. Closed-form power allocation and maximum weight bipartite matching are integrated to enable efficient user pairing and resource management. Simulation results show that the proposed scheme significantly enhances system performance in terms of throughput, spectral efficiency, and interference management, while enabling robustness against channel uncertainties in dynamic IIoT environments. The findings indicate that combining model-free reinforcement learning with conventional optimization provides a viable solution for adaptive resource management in dynamic UAV-GAP cooperative communication scenarios. Full article

In scalable video transmission research, the video transmission process is commonly modeled as a Markov decision process, where deep reinforcement learning (DRL) methods are employed to optimize the wireless transmission of scalable videos. Furthermore, the adaptive DRL algorithm can address the energy shortage problem caused by the uncertainty of energy capture and accumulated storage, thereby reducing video interruptions and enhancing user experience. To further optimize resources in wireless energy transmission and tackle the challenge of balancing exploration and exploitation in the DRL algorithm, this paper develops an adaptive DRL algorithm that extends classical DRL frameworks by integrating dropout techniques during both the training and prediction processes. Moreover, to address the issue of continuous negative rewards, which are often attributed to incomplete training in the wireless video transmission DRL algorithm, this paper introduces the Cramér large deviation principle for specific discrimination. It identifies the optimal negative reward frequency boundary and minimizes the probability of misjudgment regarding continuous negative rewards. Finally, experimental validation is performed using the 2048-game environment that simulates wireless scalable video transmission conditions. The results demonstrate that the adaptive DRL algorithm described in this paper achieves superior convergence speed and higher cumulative rewards compared to the classical DRL approaches. Full article

Downhole kick is one of the most severe safety hazards in deep and ultra-deep well drilling operations. Traditional monitoring methods, which rely on surface flow rate and fluid level changes, are limited by their delayed response and insufficient sensitivity, making them inadequate for early warning. This study proposes a real-time monitoring technique for gas content in drilling fluid based on the attenuation principle of Ba-133 γ-rays. By integrating laboratory static/dynamic experiments and Geant4-11.2 Monte Carlo simulations, the influence mechanism of gas–liquid two-phase media on γ-ray transmission characteristics is systematically elucidated. Firstly, through a comparative analysis of radioactive source parameters such as Am-241 and Cs-137, Ba-133 (main peak at 356 keV, half-life of 10.6 years) is identified as the optimal downhole nuclear measurement source based on a comparative analysis of penetration capability, detection efficiency, and regulatory compliance. Compared to alternative sources, Ba-133 provides an optimal energy range for detecting drilling fluid density variations, while also meeting exemption activity limits (1 × 10⁶ Bq) for field deployment. Subsequently, an experimental setup with drilling fluids of varying densities (1.2–1.8 g/cm³) is constructed to quantify the inverse square attenuation relationship between source-to-detector distance and counting rate, and to acquire counting data over the full gas content range (0–100%). The Monte Carlo simulation results exhibit a mean relative error of 5.01% compared to the experimental data, validating the physical correctness of the model. On this basis, a nonlinear inversion model coupling a first-order density term with a cubic gas content term is proposed, achieving a mean absolute percentage error of 2.3% across the full range and R² = 0.999. Geant4-based simulation validation demonstrates that this technique can achieve a measurement accuracy of ±2.5% for gas content within the range of 0–100% (at a 95% confidence interval). The anticipated field accuracy of ±5% is estimated by accounting for additional uncertainties due to temperature effects, vibration, and mud composition variations under downhole conditions, significantly outperforming current surface monitoring methods. This enables the high-frequency, high-precision early detection of kick events during the shut-in period. The present study provides both theoretical and technical support for the engineering application of nuclear measurement techniques in well control safety. Full article

To address the challenges of high complexity in modeling the correlation of multi-dimensional stochastic variables and the difficulty of solving long-term scheduling models in continuous action spaces in multi-energy complementary systems, this paper proposes a long-term optimization scheduling method based on Deep Deterministic Policy Gradient (DDPG). First, an improved C-Vine Copula model is used to construct the multi-dimensional joint probability distribution of water, wind, and solar energy, and Latin Hypercube Sampling (LHS) is employed to generate a large number of water–wind–solar coupling scenarios, effectively reducing the model’s complexity. Then, a long-term optimization scheduling model is established with the goal of maximizing the absorption of clean energy, and it is converted into a Markov Decision Process (MDP). Next, the DDPG algorithm is employed with a noise dynamic adjustment mechanism to optimize the policy in continuous action spaces, yielding the optimal long-term scheduling strategy for the water–wind–solar multi-energy complementary system. Finally, using a water–wind–solar integrated energy base as a case study, comparative analysis demonstrates that the proposed method can improve the renewable energy absorption capacity and the system’s power generation efficiency by accurately quantifying the uncertainties of water, wind, and solar energy and precisely controlling the continuous action space during the scheduling process. Full article

Nanoimprint lithography (NIL) has emerged as a promising sub-10 nm patterning at low cost; yet, robust process control remains difficult because of time-consuming physics-based simulators and labeled SEM data scarcity. We propose a data-efficient, two-stage deep-learning framework here that directly reconstructs post-imprint SEM images from binary design layouts and delivers calibrated pixel-by-pixel uncertainty simultaneously. First, a shallow U-Net is trained on conformalized quantile regression (CQR) to output 90% prediction intervals with statistically guaranteed coverage. Moreover, per-level errors on a small calibration dataset are designed to drive an outlier-weighted and encoder-frozen transfer fine-tuning phase that refines only the decoder, with its capacity explicitly focused on regions of spatial uncertainty. On independent test layouts, our proposed fine-tuned model significantly reduces the mean absolute error (MAE) from 0.0365 to 0.0255 and raises the coverage from 0.904 to 0.926, while cutting the labeled data and GPU time by 80% and 72%, respectively. The resultant uncertainty maps highlight spatial regions associated with error hotspots and support defect-aware optical proximity correction (OPC) with fewer guard-band iterations. Extending the current perspective beyond OPC, the innovatively model-agnostic and modular design of the pipeline here allows flexible integration into other critical stages of the semiconductor manufacturing workflow, such as imprinting, etching, and inspection. In these stages, such predictions are critical for achieving higher precision, efficiency, and overall process robustness in semiconductor manufacturing, which is the ultimate motivation of this study. Full article

In nature, organisms often rely on the integration of local sensory information and prior experience to flexibly adapt to complex and dynamic environments, enabling efficient path selection. This bio-inspired mechanism of perception and behavioral adjustment provides important insights for path planning in mobile robots operating under uncertainty. In recent years, the introduction of deep reinforcement learning (DRL) has empowered mobile robots to autonomously learn navigation strategies through interaction with the environment, allowing them to identify obstacle distributions and perform path planning even in unknown scenarios. To further enhance the adaptability and path planning performance of robots in complex environments, this paper develops a deep reinforcement learning framework based on the Soft Actor–Critic (SAC) algorithm. First, to address the limited adaptability of existing transfer learning methods, we propose an action-level fusion mechanism that dynamically integrates prior and current policies during inference, enabling more flexible knowledge transfer. Second, a bio-inspired radar perception optimization method is introduced, which mimics the biological mechanism of focusing on key regions while ignoring redundant information, thereby enhancing the expressiveness of sensory inputs. Finally, a reward function based on ineffective behavior recognition is designed to reduce unnecessary exploration during training. The proposed method is validated in both the Gazebo simulation environment and real-world scenarios. Experimental results demonstrate that the approach achieves faster convergence and superior obstacle avoidance performance in path planning tasks, exhibiting strong transferability and generalization across various obstacle configurations. Full article

Developing and identifying the correct reactor model for a reaction system characterized by a high number of reaction pathways and flow regimes can be challenging. In this work, artificial neural networks (ANNs), used in deep learning, are used to develop a hybrid modelling framework for physics-based model discovery in reactions systems. The model discovery accuracy of the framework is investigated considering kinetic model parametric uncertainty, noise level, features in the data structure and experimental design optimization via a differential evolution algorithm (DEA). The hydrodynamic behaviours of both a continuously stirred tank reactor and a plug flow reactor and rival chemical kinetics models are combined to generate candidate physics-based models to describe a benzoic acid esterification synthesis in a rotating cylindrical reactor. ANNs are trained and validated from in silico data simulated by sampling the parameter space of the physics-based models. Results show that, when monitored using test data classification accuracy, ANN performance improved when the kinetic parameters uncertainty decreased. The performance improved further by increasing the number of features in the data set, optimizing the experimental design and decreasing the measurements error (low noise level). Full article

Secure communication is critically dependent on high-speed and high-security quantum random number generation (QRNG). In this work, we present a responsive approach to enhance the efficiency and security of QRNG by leveraging polarization-controlled heterodyne detection to simultaneously measure the quadrature amplitude and phase fluctuations of vacuum shot noise. To address the practical non-idealities inherent in QRNG systems, we investigate the critical impacts of imbalanced heterodyne detection, amplitude–phase overlap, finite-size effects, and security parameters on quantum conditional min-entropy derived from the entropy uncertainty principle. It effectively mitigates the overestimation of randomness and fortifies the system against potential eavesdropping attacks. For a high-security parameter of ${10}^{-20}$, QRNG achieves a true random bit extraction ratio of 83.16% with a corresponding real-time speed of 37.25 Gbps following a 16-bit analog-to-digital converter quantization and 1.4 GHz bandwidth extraction. Furthermore, we develop a deep convolutional neural network for rapid and accurate entropy evaluation. The entropy evaluation of 13,473 sets of quadrature data is processed in 68.89 s with a mean absolute percentage error of 0.004, achieving an acceleration of two orders of magnitude in evaluation speed. Extracting the shot noise with full detection bandwidth, the generation rate of QRNG using dual-quadrature heterodyne detection exceeds 85 Gbps. The research contributes to advancing the practical deployment of QRNG and expediting rapid entropy assessment. Full article