Search Results (363)

Geographic Named Entity Recognition (Geo-NER) is a crucial task for extracting geography-related entities from unstructured text, and it plays an essential role in geographic information extraction and spatial semantic understanding. Traditional approaches typically treat Geo-NER as a sequence labeling problem, where each token is assigned a single label. However, this formulation struggles to handle nested entities effectively. To overcome this limitation, we propose Geo-MRC, an improved model based on a Machine Reading Comprehension (MRC) framework that reformulates Geo-NER as a question-answering task. The model identifies entities by predicting their start positions, end positions, and lengths, enabling precise detection of overlapping and nested entities. Specifically, it constructs a unified input sequence by concatenating a type-specific question (e.g., “What are the location names in the text?”) with the context. This sequence is encoded using BERT, followed by feature extraction and fusion through Gated Recurrent Units (GRU) and multi-scale 1D convolutions, which improve the model’s sensitivity to both multi-level semantics and local contextual information. Finally, a feed-forward neural network (FFN) predicts whether each token corresponds to the start or end of an entity and estimates the span length, allowing for dynamic inference of entity boundaries. Experimental results on multiple public datasets demonstrate that Geo-MRC consistently outperforms strong baselines, with particularly significant gains on datasets containing nested entities. Full article

Background: Current fMRI (functional magnetic resonance imaging)-driven brain information decoding for visual image reconstruction techniques faces issues such as poor structural fidelity, inadequate model generalization, and unnatural visual image reconstruction in complex scenarios. Methods: To address these challenges, this study proposes a Dynamic Confidence Bayesian Adaptive Network (DCBAN). In this network model, deep nested Singular Value Decomposition is introduced to embed low-rank constraints into the deep learning model layers for fine-grained feature extraction, thus improving structural fidelity. The proposed Bayesian Adaptive Fractional Ridge Regression module, based on singular value space, dynamically adjusts the regularization parameters, significantly enhancing the decoder’s generalization ability under complex stimulus conditions. The constructed Dynamic Confidence Adaptive Diffusion Model module incorporates a confidence network and time decay strategy, dynamically adjusting the semantic injection strength during the generation phase, further enhancing the details and naturalness of the generated images. Results: The proposed DCBAN method is applied to the NSD, outperforming state-of-the-art methods by 8.41%, 0.6%, and 4.8% in PixCorr (0.361), Incep (96.0%), and CLIP (97.8%), respectively, achieving the current best performance in both structural and semantic fMRI visual image reconstruction. Conclusions: The DCBAN proposed in this thesis offers a novel solution for reconstructing visual images from fMRI signals, significantly enhancing the robustness and generative quality of the reconstructed images. Full article

The increasing deployment of islanded microgrids in disaster-prone and infrastructure-constrained regions has elevated the importance of resilient energy storage systems capable of supporting autonomous operation. Grid-forming energy storage (GFES) units—designed to provide frequency reference, voltage regulation, and black-start capabilities—are emerging as critical assets for maintaining both energy adequacy and dynamic stability in isolated environments. However, conventional storage planning models fail to capture the interplay between uncertain renewable generation, time-coupled operational constraints, and control-oriented performance metrics such as virtual inertia and voltage ride-through. To address this gap, this paper proposes a novel distributionally robust optimization (DRO) framework that jointly optimizes the siting and sizing of GFES under renewable and load uncertainty. The model is grounded in Wasserstein-metric DRO, allowing worst-case expectation minimization over an ambiguity set constructed from empirical historical data. A multi-period convex formulation is developed that incorporates energy balance, degradation cost, state-of-charge dynamics, black-start reserve margins, and stability-aware constraints. Frequency sensitivity and voltage compliance metrics are explicitly embedded into the optimization, enabling control-aware dispatch and resilience-informed placement of storage assets. A tractable reformulation is achieved using strong duality and solved via a nested column-and-constraint generation algorithm. The framework is validated on a modified IEEE 33-bus distribution network with high PV penetration and heterogeneous demand profiles. Case study results demonstrate that the proposed model reduces worst-case blackout duration by 17.4%, improves voltage recovery speed by 12.9%, and achieves 22.3% higher SoC utilization efficiency compared to deterministic and stochastic baselines. Furthermore, sensitivity analyses reveal that GFES deployment naturally concentrates at nodes with high dynamic control leverage, confirming the effectiveness of the control-informed robust design. This work provides a scalable, data-driven planning tool for resilient microgrid development in the face of deep temporal and structural uncertainty. Full article

With the increasing penetration of distributed energy resources (DERs) in distribution networks, traditional passive distribution systems are evolving into active and flexible systems capable of participating in the transmission-level energy market. Integrating distribution networks into a transmission-centric market-clearing model introduces challenges, such as capturing internal operational constraints and reflecting the economic features of distribution systems. To this end, this paper proposes a market integration method for distribution networks based on a feasible region and an accompanying bidding strategic bidding method to enable their efficient participation in the transmission-level electricity market. With a two-stage adaptive robust optimization framework, the feasible region that preserves operational characteristics of the distribution system and ensures the satisfaction of operational constraints within the distribution system is first depicted. The feasible region appears as time-coupled box-shaped regions. On this basis, a strategic bidding method is proposed based on the nested segmentation of the feasible region, jointly considering power and reserve. With it, the bidding prices of energy and reserve can be prepared, and then, together with the feasible region, can be smoothly integrated into the transmission-level market model. Numerical case studies demonstrate the effectiveness of the proposed method. Full article

Embodied energy flows link production systems with the energy sector, reflecting dependencies and structural risks under globalization and regional coordination. Guangdong, China’s most manufacturing-intensive, open, and energy-consuming province, is a central hub in both global value chains and domestic production networks, playing a pivotal role in national energy security. Understanding Guangdong’s embodied energy flows is essential for revealing the transmission of energy across multi-level spatial systems and the resilience of China’s energy infrastructure. This study integrates international (EXIOBASE) and Chinese inter-provincial input–output data to build a province-level nested global MRIO model, combined with Structural Path Analysis (SPA), to characterize Guangdong’s manufacturing embodied energy flows in domestic and international dual circulation from 2002 to 2017. Our findings confirm Guangdong’s pivotal bridging role in embodied energy transfers. First, flows are dual-directional and dominated by international transfers. Second, energy efficiency has improved, narrowing the intensity gap between export- and domestic-oriented industries. Third, flows have diversified spatially from concentration in developed regions toward developing regions, with domestic inter-provincial flows more dispersed. Finally, embodied energy remains highly concentrated across sectors, with leading industries shifting from labor- and capital-intensive to capital- and technology-intensive sectors. This research offers vital empirical evidence and policy reference for enhancing national energy security and optimizing spatial energy allocation. Full article

In graph theory and network design, the minimum cut is a fundamental measure of system connectivity and communication capacity. While prior research has largely focused on computing the minimum cut for a fixed source–sink pair, practical scenarios such as data center communication often demand a different objective: identifying the source node whose minimum cut to a designated sink is maximized. This task, which we term the Global Maximum Minimum Cut with Fixed Sink (GMMC-FS) problem, captures the goal of locating a high-capacity source relative to a shared sink node that aggregates multiple servers. The problem is of significant engineering importance, yet it is computationally challenging as it involves a nested max–min optimization. In this paper, we present a recursive reduction (RR) algorithm for solving the GMMC-FS problem. The key idea is to iteratively select pivot nodes, compute their minimum cuts with respect to the sink, and prune dominated candidates whose cut values cannot exceed that of the pivot. By recursively applying this elimination process, RR dramatically reduces the number of max-flow computations required while preserving exact correctness. Compared with classical contraction-based and Gomory–Hu tree approaches that rely on global cut enumeration, the proposed RR framework offers a more direct and scalable mechanism for identifying the source that maximizes the minimum cut to a fixed sink. Its novelty lies in exploiting the structural properties of the sink side of suboptimal cuts, which leads to both theoretical efficiency and empirical robustness across large-scale networks. We provide a rigorous theoretical analysis establishing both correctness and complexity bounds, and we validate the approach through extensive experiments. Results demonstrate that RR consistently achieves optimal solutions while significantly outperforming baseline methods in runtime, particularly on large and dense networks. Full article

A nested semi-tube hollow-core anti-resonant fiber (HC-ARF) that can support the high-purity transmission of a few polarization-maintaining modes is designed in this paper. By employing bi-thickness hybrid silica/silicon anti-resonant tubes, the birefringence of the orthogonal polarized modes is significantly improved, and the weak coupling condition of the five lowest-order polarization maintaining modes, including the LP_{01_x}, LP_{01_y}, LP_{11a_x}, LP_{11b_x}, and LP_{11a_y}, can be met. The effective refractive index difference between each pair of the supported adjacent modes is larger than 1.0 × 10⁻⁴. With hybrid multi-layer nested semi-tubes, the confinement losses of the supported modes are all less than 1.50 × 10⁻¹ dB/m within a transmission band from 1.530 to 1.620 μm. The minimum confinement losses of the LP_{01_y}, LP_{01_x}, LP_{11a_y}, LP_{11a_x}, and LP_{11b_x} modes are 3.71 × 10⁻⁴ dB/m, 1.61 × 10⁻³ dB/m, 2.00 × 10⁻² dB/m, 1.30 × 10⁻¹ dB/m, and 4.20 × 10⁻² dB/m, respectively. Meanwhile, the unwanted higher-order modes are filtered out well to guarantee the modal purity. The minimum higher-order-mode extinction ratio of the lowest-loss LP₂₁ mode to the highest-loss LP₁₁ mode remains larger than 139 from 1.545 to 1.615 μm. The numerical results highlight the potential of the proposed polarization-maintaining few-mode hollow-core anti-resonant fibers in many application fields, such as short-range and high-capacity data transmission networks, fiber sensing systems, quantum communication systems, and so on. Full article

Ship detection in synthetic aperture radar (SAR) remote sensing imagery is of great significance in military and civilian applications. However, two factors limit detection performance: (1) a high prevalence of small-scale ship targets with limited information content and (2) interference affecting ship detection from speckle noise and land–sea clutter. To address these challenges, we propose a novel end-to-end (E2E) transformer-based SAR ship detection framework, called Flow-Aligned Nested Transformer for SAR Small Ship Detection (FANT-Det). Specifically, in the feature extraction stage, we introduce a Nested Swin Transformer Block (NSTB). The NSTB employs a two-level local self-attention mechanism to enhance fine-grained target representation, thereby enriching features of small ships. For multi-scale feature fusion, we design a Flow-Aligned Depthwise Efficient Channel Attention Network (FADEN). FADEN achieves precise alignment of features across different resolutions via semantic flow and filters background clutter through lightweight channel attention, further enhancing small-target feature quality. Moreover, we propose an Adaptive Multi-scale Contrastive Denoising (AM-CDN) training paradigm. AM-CDN constructs adaptive perturbation thresholds jointly determined by a target scale factor and a clutter factor, generating contrastive denoising samples that better match the physical characteristics of SAR ships. Finally, extensive experiments on three widely used open SAR ship datasets demonstrate that the proposed method achieves superior detection performance, outperforming current state-of-the-art (SOTA) benchmarks. Full article

Rice, as one of the world’s four major staple crops, is frequently threatened by pests and diseases during its growth. With the rapid expansion of agricultural information data, the effective management and utilization of such data have become crucial for the development of agricultural informatization. Named entity recognition technology offers precise support for the early prevention and control of crop pests and diseases. However, entity recognition for rice pests and diseases faces challenges such as structural complexity and prevalent nesting issues. Inspired by biological visual mechanisms, we propose a deep learning model capable of extracting multi-granularity features. Text representations are encoded using BERT, and the model enhances its ability to capture nested boundary information through multi-granularity convolutional neural networks (CNNs). Finally, sequence modeling and labeling are performed using a bidirectional long short-term memory network (BiLSTM) combined with a conditional random field (CRF). Experimental results demonstrate that the proposed model effectively identifies entities related to rice diseases and pests, achieving an F1 score of 91.74% on a self-constructed dataset. Full article

In recent years, frequent typhoon-induced disasters have significantly increased the risk of power grid outages, posing severe challenges to the secure and stable operation of distribution grids with high penetration of distributed photovoltaic (PV) systems. Furthermore, during post-disaster recovery, the dual uncertainties of distributed PV output and the charging/discharging behavior of flexible resources such as electric vehicles (EVs) complicate the configuration and scheduling of mobile energy storage systems (MESS). To address these challenges, this paper proposes a two-stage robust optimization framework for dynamic recovery of distribution grids: Firstly, a multi-stage decision framework is developed, incorporating MESS site selection, network reconfiguration, and resource scheduling. Secondly, a spatiotemporal coupling model is designed to integrate the dynamic dispatch behavior of MESS with the temporal and spatial evolution of disaster scenarios, enabling dynamic path planning. Finally, a nested column-and-constraint generation (NC&CG) algorithm is employed to address the uncertainties in PV output intervals and EV demand fluctuations. Simulations on the IEEE 33-node system demonstrate that the proposed method improves grid resilience and economic efficiency while reducing operational risks. Full article

Assimilation of weather radar data, as well as additional data like lightning data, in high-resolution weather forecast models is a promising method to improve short-term forecasting (nowcasting) of flash-flood events. A data assimilation system based on the Weather Research and Forecasting model is used in this study, with radar reflectivity and radial velocity data collected with X-band Doppler polarimetric radar in the area of Athens, Greece, and lightning observations obtained from a lightning detection network covering Greece. Radar data are assimilated with the four-dimensional variational method, which includes a full-hydrometeor assimilation scheme, in a nested domain of the model with a resolution of 3 km. Humidity, vertical velocity and horizontal wind divergence profiles estimated from lightning data are assimilated with a three-dimensional variation method in the parent domain of the model with a resolution of 9 km. The results from a case study are presented to show the effect of assimilating each type of data. Full article

Bayesian analysis is particularly useful for inferring models and their parameters given data. This is a common task in metabolic modeling, where models of varying complexity are used to interpret data. Nested sampling is a class of probabilistic inference algorithms that are particularly effective for estimating evidence and sampling the parameter posterior probability distributions. However, the practicality of nested sampling for metabolic network inference has yet to be studied. In this technical report, we explore the amalgamation of nested sampling, specifically diffusive nested sampling, with reversible jump Markov chain Monte Carlo. We apply the algorithm to two synthetic problems from the field of metabolic flux analysis. We present run times and share insights into hyperparameter choices, providing a useful point of reference for future applications of nested sampling to metabolic flux problems. Full article

Named entity recognition (NER) is one of the fundamental tasks in building knowledge graphs. For some domain-specific corpora, the text descriptions exhibit limited standardization, and some entity structures have entity nesting. The existing entity recognition methods have problems such as word matching noise interference and difficulty in distinguishing different entity labels for the same character in sequence label prediction. This paper proposes a span-based feature reuse stacked bidirectional long short term memory network (BiLSTM) nested named entity recognition (SFRSN) model, which transforms the entity recognition of sequence prediction into the problem of entity span suffix category classification. Firstly, character feature embedding is generated through bidirectional encoder representation of transformers (BERT). Secondly, a feature reuse stacked BiLSTM is proposed to obtain deep context features while alleviating the problem of deep network degradation. Thirdly, the span feature is obtained through the dilated convolution neural network (DCNN), and at the same time, a single-tail selection function is introduced to obtain the classification feature of the entity span suffix, with the aim of reducing the training parameters. Fourthly, a global feature gated attention mechanism is proposed, integrating span features and span suffix classification features to achieve span suffix classification. The experimental results on four Chinese-specific domain datasets demonstrate the effectiveness of our approach: SFRSN achieves micro-F1 scores of 83.34% on ontonotes, 73.27% on weibo, 96.90% on resume, and 86.77% on the supply chain management dataset. This represents a maximum improvement of 1.55%, 4.94%, 2.48%, and 3.47% over state-of-the-art baselines, respectively. The experimental results demonstrate the effectiveness of the model in addressing nested entities and entity label ambiguity issues. Full article

Face recognition under uneven illumination conditions presents significant challenges, as asymmetric shadows often obscure facial features while overexposed regions lose critical texture details. To address this problem, a novel symmetry-aware illumination enhancement method named face shadow detection network (FSDN) is proposed, which features a nested U-Net architecture combined with Gaussian convolution. This method enables precise illumination intensity maps for the given face images through higher-order quadratic enhancement curves, effectively extending the low-light dynamic range while preserving essential facial symmetry. Comprehensive evaluations on the Extended Yale B and CMU-PIE datasets demonstrate the superiority of the proposed FSDN over conventional approaches, achieving structural similarity (SSIM) indices of 0.48 and 0.59, respectively, along with remarkably low face recognition error rates of 1.3% and 0.2%, respectively. The key innovation of this work lies in its simultaneous optimization of illumination uniformity and facial symmetry preservation, thereby significantly improving face analysis reliability under challenging lighting conditions. Full article

Chronic myeloid leukemia (CML) is a clonal myeloproliferative disorder caused by the BCR::ABL1 fusion gene, resulting from a reciprocal translocation between chromosomes 22 and 9. Quantification of BCR::ABL1 transcript levels in peripheral blood by RT-qPCR represents the gold standard for molecular response (MR) monitoring, providing essential clinical information on treatment efficacy. Xpert^® BCR-ABL Ultra is a fully automated in vitro diagnostic test that quantitatively detects e13a2 and e14a2 BCR::ABL1 transcripts using a single-use cartridge that integrates RNA extraction, cDNA synthesis, nested real-time PCR, and signal detection within a rapid, closed, and user-friendly system. In this study, we evaluated Xpert^® BCR-ABL Ultra as an alternative to validated systems currently used by four highly specialized Italian laboratories affiliated with the Italian national laboratory network for CML. A total of 129 peripheral blood samples from CML patients at various disease stages, along with two external quality control materials, were analyzed. We assessed the test’s repeatability, specificity, and stability. Concordance of BCR::ABL1%IS values generated by the different methods was evaluated using EUTOS criteria and Bland–Altman analysis. Finally, MR value concordance was analyzed based on European LeukemiaNet recommendations or calculated using the formula 2 − log₁₀(BCR::ABL1%IS). Xpert^® BCR-ABL Ultra demonstrated high repeatability and stability. The BCR::ABL1%IS values obtained with this assay showed strong concordance with those generated by local reference methods, and MR classifications were consistent across platforms. These findings confirm the robustness, accuracy, and efficiency of the Xpert^® BCR-ABL Ultra assay, supporting its use as a reliable alternative to currently validated systems for the routine clinical monitoring of CML patients. Full article