Search Results (246)

To achieve a complete circular economy for used electric vehicle batteries, it is essential to implement a disassembly step. Given the significant diversity of battery geometries and designs, a high degree of flexibility is required for automated disassembly processes. The incorporation of human–robot interaction provides a valuable degree of flexibility in the process workflow. However, human behavior is characterized by unpredictable timing and variable task durations, which add considerable complexity to process planning. Therefore, it is crucial to develop a robust strategy for coordinating human and robotic tasks to manage the scheduling of production activities efficiently. This study proposes a global optimization approach to the scheduling of production activities, which employs a genetic algorithm with the objective of minimizing the total production time while simultaneously reducing the idle time of both the human operator and robot. The proposed approach is concerned with optimizing the sequencing of disassembly tasks, considering both temporal and exclusion constraints, to guarantee that tasks deemed hazardous are not executed in the presence of a human. This approach is based on a two-level adaptation framework developed in RoboDK (Robot Development Kit, v5.4.3.22231, 2022, RoboDK Inc., Montréal, QC Canada). At the first level, offline optimization is performed using a genetic algorithm to determine the optimal task sequencing strategy. This stage anticipates human behavior by proposing disassembly sequences aligned with expected human availability. At the second level, an online reactive adjustment refines the plan in real time, adapting it to actual human interventions and compensating for deviations from initial forecasts. The effectiveness of this global optimization strategy is evaluated against a non-global approach, in which the problem is partitioned into independent subproblems solved separately and then integrated. The results demonstrate the efficacy of the proposed approach in comparison with a non-global approach, particularly in scenarios where humans arrive earlier than anticipated. Full article

Transformer-based models have accelerated EEG motor imagery (MI) decoding by using self-attention to capture long-range temporal structures while complementing spatial inductive biases. This systematic survey of Scopus-indexed works from 2020 to 2025 indicates that reported advances are concentrated in offline, protocol-heterogeneous settings; inconsistent preprocessing, non-standard data splits, and sparse efficiency frequently reporting cloud claims of generalization and real-time suitability. Under session- and subject-aware evaluation on the BCIC IV 2a/2b dataset, typical performance clusters are in the high-80% range for binary MI and the mid-70% range for multi-class tasks with gains of roughly 5–10 percentage points achieved by strong hybrids (CNN/TCN–Transformer; hierarchical attention) rather than by extreme figures often driven by leakage-prone protocols. In parallel, transformer-driven denoising—particularly diffusion–transformer hybrids—yields strong signal-level metrics but remains weakly linked to task benefit; denoise → decode validation is rarely standardized despite being the most relevant proxy when artifact-free ground truth is unavailable. Three priorities emerge for translation: protocol discipline (fixed train/test partitions, transparent preprocessing, mandatory reporting of parameters, FLOPs, per-trial latency, and acquisition-to-feedback delay); task relevance (shared denoise → decode benchmarks for MI and related paradigms); and adaptivity at scale (self-supervised pretraining on heterogeneous EEG corpora and resource-aware co-optimization of preprocessing and hybrid transformer topologies). Evidence from subject-adjusting evolutionary pipelines that jointly tune preprocessing, attention depth, and CNN–Transformer fusion demonstrates reproducible inter-subject gains over established baselines under controlled protocols. Implementing these practices positions transformer-driven BCIs to move beyond inflated offline estimates toward reliable, real-time neurointerfaces with concrete clinical and assistive relevance. Full article

The deployment of Federated Learning (FL) in edge environments is often impeded by system heterogeneity, non-independent and identically distributed (non-IID) data, and constrained communication resources, which collectively hinder training efficiency and scalability. To address these challenges, this paper presents FLEX-SFL, a flexible and efficient split federated learning framework that jointly optimizes model partitioning, client selection, and communication scheduling. FLEX-SFL incorporates three coordinated mechanisms: a device-aware adaptive segmentation strategy that dynamically adjusts model partition points based on client computational capacity to mitigate straggler effects; an entropy-driven client selection algorithm that promotes data representativeness by leveraging label distribution entropy; and a hierarchical local asynchronous aggregation scheme that enables asynchronous intra-cluster and inter-cluster model updates to improve training throughput and reduce communication latency. We theoretically establish the convergence properties of FLEX-SFL under convex settings and analyze the influence of local update frequency and client participation on convergence bounds. Extensive experiments on benchmark datasets including FMNIST, CIFAR-10, and CIFAR-100 demonstrate that FLEX-SFL consistently outperforms state-of-the-art FL and split FL baselines in terms of model accuracy, convergence speed, and resource efficiency, particularly under high degrees of statistical and system heterogeneity. These results validate the effectiveness and practicality of FLEX-SFL for real-world edge intelligent systems. Full article

Objective: This study evaluated the cost-effectiveness of sacituzumab govitecan (SG) compared with single-agent chemotherapy of the physician’s choice (TPC) from the perspective of Taiwan’s National Health Insurance. Methods: A partitioned survival model was developed to assess outcomes in patients with metastatic triple-negative breast cancer (mTNBC). Clinical data were derived from the ASCENT trial, while direct medical costs were obtained from Taiwan’s National Health Insurance Administration (NHIA). Utility values were taken from published literature. The primary outcome was the incremental cost-effectiveness ratio (ICER), expressed as cost per quality-adjusted life year (QALY) gained. One-way and probabilistic sensitivity analyses were performed to examine parameter uncertainty and test the robustness of the results. Results: In the base-case analysis, SG was associated with an incremental cost of USD 121,836 per QALY gained—exceeding Taiwan’s willingness-to-pay (WTP) threshold of USD 102,120. One-way sensitivity analyses indicated that SG drug cost was the primary driver of ICER variability. Probabilistic sensitivity analysis showed that reducing the price of SG by 50% increased the likelihood of cost-effectiveness. Conclusions: From the NHIA perspective, SG is not cost-effective for patients with advanced or metastatic TNBC at its current price. Substantial price reductions would be required for SG to become cost-effective under the WTP threshold of USD 102,120 per QALY. Full article

Multi-modal image segmentation is a key task in various fields such as urban planning, infrastructure monitoring, and environmental analysis. However, it remains challenging due to complex scenes, varying object scales, and the integration of heterogeneous data sources (such as RGB, depth maps, and infrared). To address these challenges, we proposed a novel multi-modal segmentation framework, DyFuseNet, which features dynamic adaptive windows and cross-scale feature fusion capabilities. This framework consists of three key components: (1) Dynamic Window Module (DWM), which uses dynamic partitioning and continuous position bias to adaptively adjust window sizes, thereby improving the representation of irregular and fine-grained objects; (2) Scale Context Attention (SCA), a hierarchical mechanism that associates local details with global semantics in a coarse-to-fine manner, enhancing segmentation accuracy in low-texture or occluded regions; and (3) Hierarchical Adaptive Fusion Architecture (HAFA), which aligns and fuses features from multiple modalities through shallow synchronization and deep channel attention, effectively balancing complementarity and redundancy. Evaluated on benchmark datasets (such as ISPRS Vaihingen and Potsdam), DyFuseNet achieved state-of-the-art performance, with mean Intersection over Union (mIoU) scores of 80.40% and 80.85%, surpassing MFTransNet by 1.91% and 1.77%, respectively. The model also demonstrated strong robustness in challenging scenes (such as building edges and shadowed objects), achieving an average F1 score of 85% while maintaining high efficiency (26.19 GFLOPs, 30.09 FPS), making it suitable for real-time deployment. This work presents a practical, versatile, and computationally efficient solution for multi-modal image analysis, with potential applications beyond remote sensing, including smart monitoring, industrial inspection, and multi-source data fusion tasks. Full article

After the original support system in the auxiliary transportation roadway of the northern wing of the Zhaoxian Mine failed, the extent of damage and deformation varied significantly across different sections of the drift. A single support method could not meet the engineering requirements. Therefore, this paper conducted research on the classification of roadway damage and zoning repair. The overall damage characteristics of the roadway are described by three indicators: roadway deformation, development of rock mass fractures, and water seepage conditions. These are further refined into nine secondary indicators. In summary, a rock mass damage combination weighting evaluation model based on the FAHP–entropy weight TOPSIS method is proposed. According to this model, the degree of damage to the roadway is divided into five grades. After analyzing the damage conditions and support requirements at each grade, corresponding zoning repair plans are formulated by adjusting the parameters of bolts, cables, channel steel beams, and grouting materials. At the same time, the reliability of partition repair is verified using FLAC3D 6.0 numerical simulation software. Field monitoring results demonstrated that this approach not only met the support requirements for the roadway but also improved the utilization rate of support materials. This provides valuable guidance for the design of support systems for roadways with similar heterogeneous damage. Full article

Background/Objectives: Sugemalimab demonstrated clinical efficacy in the GEMSTONE-303 trial, but its cost-effectiveness remains unclear. This study aims to evaluate the cost-effectiveness of sugemalimab in combination with chemotherapy (CAPOX) as a first-line treatment for patients with advanced or metastatic gastric or gastroesophageal junction (G/GEJ) adenocarcinoma, compared to chemotherapy alone, from the perspective of Taiwan’s healthcare payer. Methods: A partitioned survival model was developed to simulate outcomes over a 40-year time horizon, and model parameters were derived from GEMSTONE-303 and the wider literature. Health benefits were measured in quality-adjusted life-years (QALYs), and only direct medical costs were included, with both discounted at an annual rate of 3%. The willingness-to-pay threshold was set at three times the 2024 GDP per capita. Deterministic and probabilistic sensitivity analyses were conducted alongside scenario analyses. Results: Compared to capecitabine and oxaliplatin (CAPOX) alone, adding sugemalimab yielded an incremental gain of 0.39 QALYs at an additional cost of USD 47,020, resulting in an incremental net monetary benefit of −USD 7478. Conclusions: Sugemalimab plus CAPOX is not cost-effective for advanced or metastatic G/GEJ adenocarcinoma from the Taiwan payer’s perspective. Achieving cost-effectiveness would require a 20–30% price reduction for sugemalimab (to USD 1204–USD 1376 per 600 mg), assuming first-line therapy is administered for the median treatment duration observed in the GEMSTONE-303 trial. If reimbursement continued until disease progression, a reduction of approximately 68% would be required (USD 550 per 600 mg). Full article

Micro-expressions are extremely subtle and short-lived facial muscle movements that often reveal an individual’s genuine emotions. However, micro-expression recognition (MER) remains highly challenging due to its short duration, low motion intensity, and the imbalanced distribution of training samples. To address these issues, this paper proposes a Global–Local Feature Fusion Network (GLFNet) to effectively extract discriminative features for MER. Specifically, GLFNet consists of three core modules: the Global Attention (LA) module, which captures subtle variations across the entire facial region; the Local Block (GB) module, which partitions the feature map into four non-overlapping regions to emphasize salient local movements while suppressing irrelevant information; and the Adaptive Feature Fusion (AFF) module, which employs an attention mechanism to dynamically adjust channel-wise weights for efficient global–local feature integration. In addition, a class-balanced loss function is introduced to replace the conventional cross-entropy loss, mitigating the common issue of class imbalance in micro-expression datasets. Extensive experiments are conducted on three benchmark databases, SMIC, CASME II, and SAMM, under two evaluation protocols. The experimental results demonstrate that under the Composite Database Evaluation protocol, GLFNet consistently outperforms existing state-of-the-art methods in overall performance. Specifically, the unweighted F1-scores on the Combined, SAMM, CASME II, and SMIC datasets are improved by 2.49%, 2.02%, 0.49%, and 4.67%, respectively, compared to the current best methods. These results strongly validate the effectiveness and superiority of the proposed global–local feature fusion strategy in micro-expression recognition tasks. Full article

Hydraulic fracturing technology significantly enhances reservoir conductivity by creating artificial fractures, serving as a crucial means for the economically viable development of low-permeability reservoirs. Accurate prediction of post-fracturing productivity is essential for optimizing fracturing parameter design and establishing scientific production strategies. However, current limitations in understanding post-fracturing production dynamics and the lack of efficient prediction methods severely constrain the evaluation of fracturing effectiveness and the adjustment of development plans. This study proposes a machine learning-based method for predicting post-fracturing productivity in multi-layer commingled production wells and validates its effectiveness using a key block from the PetroChina North China Huabei Oilfield Company. During the data preprocessing stage, the three-sigma rule, median absolute deviation, and density-based spatial clustering of applications with noise were employed to detect outliers, while missing values were imputed using the K-nearest neighbors method. Feature selection was performed using Pearson correlation coefficient and variance inflation factor, resulting in the identification of twelve key parameters as input features. The coefficient of determination served as the evaluation metric, and model hyperparameters were optimized using grid search combined with cross-validation. To address the multi-layer commingled production challenge, seven distinct datasets incorporating production parameters were constructed based on four geological parameter partitioning methods: thickness ratio, porosity–thickness product ratio, permeability–thickness product ratio, and porosity–permeability–thickness product ratio. Twelve machine learning models were then applied for training. Through comparative analysis, the most suitable productivity prediction model for the block was selected, and the block’s productivity patterns were revealed. The results show that after training with block-partitioned data, the accuracy of all models has improved; further stratigraphic subdivision based on block partitioning has led the models to reach peak performance. However, data volume is a critical limiting factor—for blocks with insufficient data, stratigraphic subdivision instead results in a decline in prediction performance. Full article

During feeder outages in the distribution network, localized power restoration using distribution resources (e.g., PVs) can ensure supply to critical loads and mitigate adverse impacts, especially when main grid support is unavailable. This study presents a power restoration strategy aiming at maximizing the restoration duration of critical loads to ensure their prioritized recovery, thereby significantly improving power system reliability. The methodology begins with load enumeration via breadth-first search (BFS) and utilizes a long short-term memory (LSTM) neural network to predict microgrid generation output. Then, an adaptive multipoint crossover genetic solving algorithm (AMCGA) is proposed, which can dynamically adjust crossover and mutation rates, enabling rapid convergence and requiring fewer parameters, thus optimizing island partitioning to prioritize critical load demands. Experimental results show that AMCGA improves convergence speed by 42.5% over the traditional genetic algorithm, resulting in longer restoration durations. Compared with other strategies that do not prioritize critical load recovery, the proposed strategy has shown superior performance in enhancing critical load restoration, optimizing island partitioning, and reducing recovery fluctuations, thereby confirming the strategy’s effectiveness in maximizing restoration and improving stability. Full article

Accurately partitioning actual evapotranspiration ET_{c act} into soil evaporation E_s and plant transpiration T_{c act} is crucial for improving water use efficiency and devising precise irrigation schedules. In water-saving irrigated rice fields, ET_{c act}, E_s and T_{c act} were estimated using a dual crop coefficient method based on three approaches: FAO56 adjusted, locally calibrated and leaf area index LAI-based coefficients. Continuous measurements of hourly and daily ET_{c act}, E_s and T_{c act} with weighing lysimeters were used to validate these coefficients. Results showed that hourly ET_{c act}, E_s and T_{c act} exhibited a distinct inverted “U” shape single-peak trend. Daily ET_{c act} and T_{c act}, along with the corresponding crop coefficients K_{c act} and basal crop coefficients K_{cb act}, initially increased and then decreased throughout the rice growth stages, while daily E_s and soil evaporation coefficient K_{e act} were high during the initial stage and gradually decreased as the development stage progressed. FAO56 adjusted coefficients consistently underestimated both hourly and daily ET_{c act}, E_s and T_{c act}. Locally calibrated basal crop coefficients K_{cb Cal} were determined as 0.28, 1.17 and 1.09 for the initial, mid-season and end-season stages, respectively, and locally calibrated turbulent transport coefficient of water vapor K_{cp Cal} (recommended as 1.2 by FAO) was determined to be 1.59. Based on these calibrated coefficients, estimates of hourly and daily evapotranspiration ET_{c Cal}, soil evaporation E_{s Cal} and plant transpiration T_{c Cal} performed poorly during the initial stage but showed improved accuracy during subsequent growth stages. Hourly and daily evapotranspiration and its components based on LAI-based coefficients exhibited similar performance in estimating measurements, albeit slightly inferior to FAO56 calibrated coefficients. Overall, both the FAO56 calibrated coefficients and LAI-based coefficients are recommended for estimating evapotranspiration and its components at daily and hourly scales. These research findings provide valuable insights for optimizing irrigation regimes and improving water use efficiency in rice cultivation. Full article

Understanding how environmental factors regulate photosynthetic energy partitioning is crucial for enhancing crop resilience in future climates. This study investigated the light-response dynamics of sweet sorghum (Sorghum bicolor L. Moench) leaves under combinations of CO₂ concentrations (250, 410, and 550 μmol mol⁻¹) and temperatures (30 °C and 35 °C), using integrated chlorophyll fluorescence measurements and mechanistic photosynthesis modeling. Our results revealed that elevating CO₂ from 250 to 550 μmol mol⁻¹ significantly increased the maximum electron transport rate (J_max) by up to 57%, and enhanced the effective light absorption cross-section (σ′_ik) by 64% under high light and elevated temperature (35 °C), indicating improved photochemical efficiency and light-harvesting capability. Concurrently, these adjustments reduced PSII down-regulation. Increased temperature stimulated thermal dissipation, reflected in a rise in non-photochemical quenching (NPQ) by 0.13–0.26 units, accompanied by a reduction in the number of excited-state pigment molecules (N_k) by 20–33%. The strongly coordinated responses between quantum yield (Φ_PSII) and σ′_ik highlight a dynamic balance among photochemistry, heat dissipation, and fluorescence. These findings elucidate the synergistic photoprotective and energy-partitioning strategies that sweet sorghum employs under combined CO₂ enrichment and heat stress, providing mechanistic insights for optimizing photosynthetic performance in C₄ crops in a changing climate. Full article

In supervised fine-tuning (SFT) for Text2SQL tasks, particularly for databases with numerous tables, encoding schema features requires excessive tokens, escalating GPU resource requirements during fine-tuning. To bridge this gap, we propose LR-SQL, a general dual-model SFT framework comprising a schema linking model and an SQL generation model. At the core of our framework lies the schema linking model, which is trained on a novel downstream task termed slice-based related table filtering. This task dynamically partitions a database into adjustable slices of tables and sequentially evaluates the relevance of each slice to the input query, thereby reducing token consumption per iteration. However, slicing fragments destroys database information, impairing the model’s ability to comprehend the complete database. Thus, we integrate Chain of Thought (CoT) in training, enabling the model to reconstruct the full database context from discrete slices, thereby enhancing inference fidelity. Ultimately, the SQL generation model uses the result from the schema linking model to generate the final SQL. Extensive experiments demonstrate that our proposed LR-SQL reduces total GPU memory usage by 40% compared to baseline SFT methods, with only a 2% drop in table prediction accuracy for the schema linking task and a negligible 0.6% decrease in overall Text2SQL Execution Accuracy. Full article

This study investigated the impact of green tea addition on microbial community dynamics during Daqu fermentation, a critical process in traditional baijiu production. Four Daqu variants (0%, 10%, 20%, 30% tea) were analyzed across six fermentation periods using 16S rRNA/ITS sequencing, coupled with STR, TDR, Sloan neutral model, and phylogenetic analyses. Results showed time-dependent increases in bacterial/fungal richness, with 30% tea maximizing species richness. Tea delayed bacterial shifts until day 15 but accelerated fungal reconstruction from day 6, expanding the temporal response window. While stochastic processes dominated initial assembly (77–94% bacteria, 88–99% fungi), deterministic processes intensified with tea concentration, particularly in fungi (1% → 12%). Tea increased bacterial dispersal limitation and reduced phylogenetic conservatism of endogenous factors. This work proposed a framework for rationally engineering fermentation ecosystems by decoding evolutionary-ecological rules of microbial assembly. It revealed how plant-derived additives can strategically adjust niche partitioning and ancestral constraints to reprogram microbiome functionality. These findings provided a theoretical foundation in practical strategies for optimizing industrial baijiu production through targeted ecological interventions. Full article

Due to increasing demand for the exploration of marine resources, underwater optical wireless sensor networks (UOWSNs) have emerged as a promising solution by offering higher bandwidth and lower latency compared to traditional underwater acoustic wireless sensor networks (UAWSNs), with their existing routing protocols facing challenges in energy consumption and packet forwarding. To address these challenges, this paper proposes an energy-partitioned routing protocol based on an advancement function (EPAR) for UOWSNs. By dynamically classifying the nodes into high-energy and low-energy ones, the proposed EPAR algorithm employs an adaptive weighting strategy to prioritize the high-energy nodes in relay selection, thereby balancing network load and extending overall lifetime. In addition, a tunable advancement function is adopted by the proposed EPAR algorithm by comprehensively considering the Euclidean distance and steering angle toward the sink node. By adjusting a tunable parameter α, the function guides forwarding decisions to ensure energy-efficient and directionally optimal routing. Additionally, by employing a hop-by-hop neighbor discovery mechanism, the proposed algorithm enables each node to dynamically update its local neighbor set, thereby improving relay selection and mitigating the impact of void regions on the packet delivery ratio (PDR). Simulation results demonstrate that EPAR can obtain up to about a 10% improvement in PDR and up to about a 30% reduction in energy depletion, with a prolonged network lifetime when compared to the typical algorithms adopted in the simulations. Full article