The ancient sea dikes of the Qiantang River have a history spanning more than 400 years since their construction in the 21st year of the Jiajing reign of the Ming Dynasty (1542). Despite centuries of exposure to wind, rain, and tidal waves, as well as extensive damage caused by modern human activities, more than 40 kilometers of these historic sea dikes remain in service on the front line of flood control, continuing to play a vital role. Owing to their long history, the ancient sea dikes have also fostered a rich body of local folk traditions and accumulated profound cultural significance for the communities along both banks of the estuary. They therefore represent not only important hydraulic engineering works but also valuable historical and cultural heritage. With the passage of time and changes in hydrological and topographical conditions, many of the surviving ancient sea dikes have developed structural problems, including fractured facing stones and detached stone blocks. Consequently, reproducing traditional glutinous rice mortar using modern techniques and restoring the ancient sea dikes in accordance with cultural heritage conservation standards is of great significance for preserving historical and cultural heritage, promoting water-related culture, and maintaining the engineering functions of these historic structures. This study focuses on the ancient fish-scale stone sea dike along the Haining section of Jiaxing. Building upon previous studies and systematic analyses of ancient mortar samples, the composition of traditional glutinous rice mortar was analyzed and optimized. Through similarity tests, compressive strength tests, and field experiments, a glutinous rice mortar suitable for preliminary application in the restoration of ancient sea dikes was successfully developed. After repeated experimental verification, the optimal formulation was determined to consist of calcium hydroxide powder, calcium carbonate powder, gypsum powder, pregelatinized glutinous rice flour, a water-reducing agent, cellulose, and alum. The prepared mortar exhibited satisfactory bonding performance in similarity tests, making the bonded stone specimens difficult to separate. In compressive strength tests, the mortar generally achieved strengths exceeding 0.7 MPa. Field tests further demonstrated good resistance to hydraulic erosion and drying-induced cracking, indicating that the developed mortar has preliminarily met the requirements for practical application in sea dike restoration projects.

This paper conducts a systematic review of artificial intelligence integration in programming, history, and English education, focusing on technological applications, practical outcomes, and existing challenges. Against the backdrop of global educational digitalization, AI technologies such as generative models, adaptive learning systems, intelligent tutoring systems, and AI-enhanced learning management platforms have addressed long-standing issues in these disciplines, including low engagement in programming, abstract historical instruction, and one-size-fits-all English teaching. By analyzing 10 recent scholarly works, this study first categorizes mainstream AI educational technologies, then compares their discipline-specific uses in code generation, historical simulation, argument feedback, language practice, and adaptive assessment. It also discusses how these tools influence learner autonomy, teacher workload, classroom interaction, and educational equity. Findings reveal that AI enhances personalized learning but requires pedagogical alignment to avoid superficial knowledge acquisition, student overreliance, data bias, and integrity risks. Future research should prioritize interdisciplinary AI education frameworks, ethical governance mechanisms, teacher training, and long-term impact assessments to promote sustainable AI integration in K-12 and higher education.

As the feature size of silicon-based field-effect transistors approaches its physical limit, the surge in leakage current and static power consumption caused by short-channel effects has become a critical obstacle to sustaining Moore's law. Two-dimensional (2D) semiconductors, with their atomically thin bodies, van der Waals interfaces, and tunable band gaps, offer new material and device strategies to overcome the power bottleneck. Centering on the subthreshold swing—the core figure of merit for low-power transistors—this article reviews the research progress of 2D semiconductors represented by transition metal dichalcogenides, black phosphorus, and tellurene from four dimensions: material systems, device structures, process integration, and application compatibility. It compares the near-ideal switching characteristics of monolayer MoS₂, the high mobility of black phosphorus, and the stability limitations of black phosphorus. It summarizes the roles of gate-all-around architectures, vertical heterojunctions, edge contacts, and wafer-scale integration in improving electrostatic control, reducing contact resistance, and enhancing manufacturability. This article argues that the key challenges for low-power transistors based on 2D semiconductors have shifted from individual performance verification to a comprehensive balance among material stability, contact interfaces, defect control, CMOS compatibility, and scenario-specific reliability.

Retrieval-Augmented Generation (RAG) improves language-model answers by retrieving external evidence before generation. However, its reliability depends on the retrieved context. In real settings, passages can be irrelevant, incomplete, conflicting, or poorly ordered. These problems may reduce accuracy and explainability. This study tests how retrieval noise affects RAG and whether reranking, citation-aware generation, and lightweight verification can improve system behaviour. A controlled experiment was conducted on a small HotpotQA subset using BM25, Sentence-BERT, and FAISS. Four systems were compared: vanilla RAG, reranking-only, citation-only, and a full enhanced system. Results show that reranking achieved the highest average noisy F1, but the gain over vanilla RAG was small. The full enhanced system achieved better faithfulness, groundedness, and citation precision, but did not improve average noisy F1. This suggests that robust and explainable RAG is a multi-objective problem.

The bandgap reference voltage source is a critical module in analog and mixed-signal integrated circuits, with its performance directly impacting the accuracy and stability of circuits such as ADCs, DACs, LDOs, power management systems, and sensor interfaces. As integrated circuits evolve toward lower voltage, lower power consumption, and higher precision, traditional bandgap references face increasing demands in terms of temperature stability, power supply rejection ratio (PSRR), and process compatibility. This paper reviews key performance enhancement methods for bandgap reference voltage sources. It begins by introducing the fundamental principles of traditional BJT-based bandgap references and fully CMOS voltage reference circuits, analyzing how CTAT and PTAT components complement each other to generate a stable reference voltage. Subsequently, recent typical reference source structures are categorized and compared across three aspects: temperature stability improvement, PSRR enhancement, and fully CMOS low-power implementation. In temperature compensation, methods such as segmented curvature compensation, exponential and logarithmic curvature compensation, resistor temperature coefficient compensation, subthreshold compensation, and base current compensation are emphasized. For PSRR improvement, techniques like pre-regulation, filtering, feedback enhancement, LDO stabilization, and transistor structure optimization are analyzed. Regarding fully CMOS voltage references, developments in MOS-only, subthreshold CMOS, 2-T structures, β-multiplier structures, and flexible substrate all-in-one reference circuits are summarized. Comparative results indicate that the design of high-performance bandgap references has shifted from single-compensation methods to joint optimization of multiple mechanisms, synergistically improving overall performance through temperature compensation, PSRR enhancement, low-power biasing, and process trimming. This review provides a reference for the design and optimization of low-temperature-drift, high-PSRR, and low-power bandgap reference voltage sources in future research.

With the development of minimally invasive surgery and intelligent medical technologies, traditional rigid surgical robots have gradually revealed limitations in complex orthopedic ligament reconstruction procedures, particularly in terms of insufficient dexterity and difficulty adapting to narrow anatomical spaces. During the processes of bone tunnel creation and ligament graft delivery, surgical instruments must possess both high flexibility to maneuver around soft tissues and sufficient stiffness to ensure precise and stable operation. Therefore, achieving dynamic switching between flexible and rigid mechanical properties has become an important research direction in the field of orthopedic surgical robotics. Based on this background, this paper explores the application potential of a biomimetic continuum robot utilizing phase change materials (PCMs) in orthopedic ligament reconstruction surgery. Through a literature review, the paper examines the current development of continuum robots and phase change materials, analyzes the dexterous operational advantages of continuum robots in minimally invasive medical procedures, and investigates the stiffness-regulation characteristics of phase change materials. Building upon these findings, a novel biomimetic continuum robot structure integrating a tensegrity framework with phase change materials is proposed. This design enables dynamic switching between a "flexible navigation" mode and a "rigid fixation" mode, thereby offering a promising solution for enhancing surgical adaptability and precision in orthopedic ligament reconstruction.

Coded matrix-vector multiplication (MVM) is often evaluated by recovery latency, but in storage-constrained edge settings the preferred operating point can change when pre-storage feasibility, materialized sparsity, amortized communication, decoding overhead, and failure modes are accounted for together. This paper studies a narrow resource-accounting question: when a per-worker storage budget cannot accommodate a fixed dense coded design, can systematic low-weight redundancy provide a feasible reliability-latency compromise for sparse MVM? The simulator separates offline encoded-block placement from online worker-level vector broadcast and reports storage feasibility, unassigned tasks, amortized communication, worker nonzero operations, source-accumulation encoding cost, materialized encoded nonzeros, decoding cost, timeout-capped latency, and failure reasons. In the configured simulator, dense coding remains latency-best when storage is unconstrained, but it requires 150.9 KiB of pre-stored blocks and about 3.07× 10 4 source-accumulation operations in the baseline setting. The low-weight design obtains 0.955 recovery success, 0.1632 s success-only mean latency, 0.2003 s timeout-capped mean latency, 68.5 KiB of pre-storage, and full feasibility after storage-constrained placement under an 8 KiB stress budget where the fixed dense and budgeted-dense placements cannot place all parity blocks. Workload, placement, support, and scaling diagnostics show that the low-weight point is useful but not optimal: placement is heuristic, support construction matters, and fixed w=3 with R/K≈1/3 weakens as K grows. The contribution is therefore a reproducible resource-accounting benchmark and lightweight reference baseline, not a new straggler-optimal or literature-faithful coded-computation scheme.

In the era of big data, with the daily generation of massive data and the wide application of data mining algorithms, accelerating data storage and processing has become an important issue that both the academic community and the industrial sector need to address. The traditional CPU and GPU hardware acceleration methods are insufficient in some scenarios that require high real-time performance and low power consumption, especially in areas such as unmanned aircraft and intelligent cameras. However, the SOC_FPGA not only integrates the rich logic resources of FPGA but also carries an ARM processor, featuring flexible design, high speed, low power consumption, and portability, which is very conducive to the rapid deployment of applications in mobile terminals. This research addresses this issue by designing a high-performance and low-power computing acceleration module for convolutional neural networks using the SOC_FPGA platform and conducted image classification experiments based on the ImageNet dataset. The experimental results show that this design can ensure that the model's classification recognition accuracy reaches over 80%, while the processing rate of input image data can reach 218.76FPS (figures per second), and the maximum system power consumption is 4.8W. Compared with other platforms, it has advantages in both acceleration effect and system power consumption.

Stock volatility forecasting is important in financial markets, as it supports risk management, portfolio allocation, derivative pricing, and investment decision-making. However, traditional statistical and time series models often struggle to capture nonlinear dynamics, cross-asset relationships, and external information influencing market behaviour. Although machine learning and deep learning methods have improved prediction performance, many existing models still rely mainly on historical market data and insufficiently incorporate financial news and investor sentiment. To address these limitations, this study proposes a hybrid Spatio-Temporal Sentiment Graph Attention Network (STSGAT), combining Long Short-Term Memory (LSTM) and Graph Neural Networks (GNNs). LSTM captures temporal dependencies within each stock, while GNN models structural relationships and information propagation across stocks. VADER-based sentiment scores from financial news are incorporated to account for both market structure and news-driven emotional dynamics. Overall, this study develops an integrated STSGAT framework combining temporal learning, graph learning, and sentiment analysis. Empirical results show that the proposed model outperforms baseline models, providing more accurate guidance for financial investment strategies and risk-related decision-making.

A digital exterior-modeling method based on unmanned aerial vehicle (UAV) oblique photogrammetry and point cloud processing is proposed. The difficulty encountered by conventional surveying methods in comprehensively acquiring spatial information from elevated and occluded areas of historic buildings is addressed by this method. The Jiuyun Fangding monument was selected as the study object. Images were acquired using a multi-altitude, multi-angle, layered circumferential flight strategy. A three-dimensional (3D) model was then reconstructed through feature matching, camera-pose estimation, multi-view stereo matching, point cloud registration, and texture mapping. Four representative dimensions were selected, and the measurements obtained from the reconstructed model were compared with field measurements. The results indicated that: (1) the absolute errors of the four dimensions ranged from 0.006 to 0.039 m, with the maximum spacing between the load-bearing columns exhibiting the largest absolute error of 0.039 m; (2) the relative errors ranged from 0.63% to 1.81%, with the width of the load-bearing column exhibiting the largest relative error of 1.81%; and (3) the reconstructed model provided a relatively complete representation of the overall architectural form and the principal structural components. The proposed method can therefore provide technical support for the digital documentation, 3D visualization, and dimensional verification of historic buildings.