Search Results (3,300)

Accurate liquid level regulation is critical for ensuring printing quality and process stability in stereolithography (SLA) 3D printing. However, traditional liquid level control methods often suffer from insufficient prediction accuracy, poor disturbance rejection capability, and limited adaptability under dynamic printing conditions. To address these challenges, this paper proposes an enhanced Transformer-based time series prediction model integrated with an xLSTM module for SLA liquid level prediction and adaptive control. By embedding the xLSTM structure into the Transformer encoder, the proposed model combines the global dependency modeling capability of self-attention mechanisms with the local temporal feature extraction capability of recurrent memory units, thereby improving the prediction accuracy and robustness of liquid level sequences. Experimental datasets were collected from an actual SLA printing platform, including multiple process-related features such as layer height, laser power, platform position, and vacuum pressure. Comparative experiments were conducted against conventional Transformer, LSTM, xLSTM, GRU, TCN, and PID-based methods. The results demonstrate that the proposed model achieves the best prediction performance, with an MAE of 0.174, RMSE of 0.222, and R² of 0.9903. Compared with the original Transformer model, the proposed approach significantly reduces prediction error and improves fitting stability. In disturbance rejection experiments, the proposed strategy effectively suppresses liquid level fluctuations under sudden pulse interference conditions, exhibiting superior robustness and dynamic response capability compared with traditional PID control. Furthermore, physical printing experiments verify that the proposed method can improve surface quality, contour accuracy, and structural stability of printed parts. Overall, the proposed Transformer–xLSTM framework provides an effective intelligent prediction and control solution for SLA liquid level regulation, offering significant potential for high-precision and intelligent additive manufacturing applications. Full article

Predictive energy management strategies (PEMSs) have attracted increasing attention in hybrid electric vehicles (HEVs) for improving fuel economy and powertrain efficiency using anticipated driving information. For PEMS, data-driven velocity prediction is widely used to capture complex driving patterns from historical trajectories and future traffic priors, but often lacks kinematic awareness, leading to physical causality violations and long-horizon state drift. To address these issues, this paper proposes a physics-informed PEMS, where a Physics-Informed Spatio-Temporal Network (PI-STN) provides control-oriented velocity information for an MPC-based energy management controller. Specifically, to address pseudo-motion in velocity prediction under standstill conditions, a global zero-speed gating mechanism is introduced; to suppress acceleration/deceleration trends that violate vehicle kinematic causality, a causal penalty is designed; and to mitigate temporal phase misalignment between data-driven predictions and physical motion priors, a Differentiable Kalman Filter (DKF) is incorporated. At each receding horizon step, the PI-STN-predicted velocity sequence is converted into future power demand through longitudinal vehicle dynamics and used by MPC for engine–battery power allocation under SOC and engine transient constraints. Under the same tested conditions, the proposed strategy reduces engine power fluctuation by 15.1% compared with BiLSTM-Transformer, and achieves an equivalent fuel consumption of 323.74 g, outperforming Transformer-KF by 3.12%. Full article

Thin-walled composite beams with unsymmetric laminates are attracting increasing attention in lightweight aerospace and mechanical structures because they enable enhanced stiffness tailoring and weight reduction beyond the limitations of conventional symmetric stacking sequences. However, despite their practical relevance, unsymmetric thin-walled laminates have received comparatively limited attention in the available buckling literature due to the additional complexity introduced by membrane–bending coupling effects. This study presents an efficient and physically transparent formulation for the buckling analysis of thin-walled composite beams with both symmetric and unsymmetric layups by combining Generalized Beam Theory (GBT) with the Ritz method. The proposed GBT-Ritz framework captures global, local, distortional, torsional, and shear-related deformation modes while consistently incorporating laminate coupling effects associated with unsymmetric configurations. The formulation is applicable to open, closed, branched, and unbranched cross-sections commonly encountered in aerospace structures. Validation against ABAQUS V2017 shell finite element models demonstrates excellent agreement (with discrepancies generally below 6%) in predicting critical buckling loads and mode shapes for various geometries and boundary conditions. The results show that unsymmetric laminates can significantly influence buckling behavior, particularly in open sections and intermediate beam lengths where coupling effects become dominant. Compared with conventional finite element approaches, the proposed method achieves substantially lower computational cost (providing speed-up factors of 1.5 to 2.5) while preserving clear physical insight into interacting instability mechanisms. Overall, the developed framework provides an efficient and practically relevant tool for the analysis and design of advanced thin-walled composite structures with tailored unsymmetric laminates. Full article

Robotic visual perception and terrain recognition are critical for autonomous locomotion and adaptive control in complex environments. However, existing models often extract weak features, confuse classes, and deliver unstable recognition. Most prior studies use end-to-end convolutional networks or single-stream feature extraction, which limits the balance between fine-grained visual representation and adaptive discrimination of confusing samples. To solve this problem, this paper proposes a vision model that blends attention mechanisms with a confusion augmentation strategy. Using an improved ResNet50 backbone, we add a local feature sharpening module and a channel–spatial attention module to strengthen edge texture and global context representation. We also design a confusion augmentation strategy based on the similarity of hard samples. It generates mixed samples through cross-perturbation in feature space, thereby improving the discrimination of highly similar terrains. Experiments show that our model achieves an accuracy of over 98.19% on various terrains, including cement, asphalt, sand, and snow. t-SNE visualization and Grad-CAM analysis demonstrate clear class separability and good interpretability, confirming the effectiveness and robustness of the approach for robotic terrain recognition. Full article

Breast cancer remains the leading cause of cancer-related deaths among women globally. Early detection through mammography is vital for improving survival rates; however, the large volume of medical images and subtle variations in lesion characteristics pose significant challenges to manual interpretation. Recent automated diagnostic models based on deep learning have shown strong potential for breast cancer classification, but challenges such as overfitting, high computational complexity, limited generalization, and insufficient interpretability remain unresolved. This paper proposes a computationally efficient and context-aware deep learning framework for breast cancer classification using transformer-based multi-scale attention mechanisms and explainable artificial intelligence (XAI). The proposed architecture integrates the Hierarchical Multi-Scale Transformer (HMT) and Edge-Aware Local Transformer (ELT) modules to jointly capture global contextual dependencies and boundary-sensitive local representations from mammographic images. ELT improves feature refinement in high-entropy regions, while HMT models global semantic interactions across multiple feature scales. In addition, an Adaptive Contextual Refinement (ACR) module is introduced to preserve semantically consistent feature representations across spatial resolutions. A Meta-Ensemble Classification (MEC) framework combining weighted SVM and K-Nearest Neighbors (KNN) classifiers is further employed using validation-guided class-adaptive weighting. The proposed framework is evaluated on four benchmark mammography datasets, namely CBIS-DDSM, DDSM, INBreast, and MIAS. The proposed model has demonstrated superior accuracy of over 99% across all breast cancer datasets. The model surpassed transformer-based baselines including Swin-T and ViT while maintaining lower parameter complexity and achieving approximately 7% higher accuracy on the CBIS-DDSM dataset. The proposed framework also demonstrated strong cross-dataset generalization and consistently achieved high precision, recall, and F1-score values across all benchmark datasets. To improve model interpretability, Grad-CAM, SHAP, Occlusion Sensitivity Analysis (OSA), and the proposed TIxAI consistency analysis framework are incorporated to provide preliminary explainability assessment for mammographic classification. The explainability analysis demonstrated spatially consistent saliency behavior across benchmark datasets; however, the current evaluation is based on internal saliency consistency rather than external clinical validation using expert lesion annotations. Overall, the proposed framework provides an effective and computationally efficient approach for automated breast cancer classification while improving model explainability and interpretability. Full article

Waterbody extraction is of great significance for water resource investigation and monitoring. In addition to RGB bands, most common satellite images have a near-infrared (NIR) band. By combining these RGB-NIR bands, certain water, vegetation, and shadow indices can be calculated. The near-infrared band and these indices are very similar to the X modality in RGB+X data (common examples include RGB-D and RGB-Thermal). However, at present, no studies have thoroughly examined multimodal feature fusion from the RGB+X perspective in order to extract waterbodies with high precision. As a result, existing algorithms do not fully utilize satellite image information and have limited generalization ability. To overcome this limitation, we propose a dual-complexity backbone for waterbody extraction from the perspective of RGB+X data semantic segmentation. Its complex Transformer branch is used to extract RGB modality features, while its simple CNN branch is used to extract X modality features. This network structure can effectively capture multimodal, global, and local features in remote sensing images. It can also fully leverage the fact that the scale of RGB image datasets in computer vision is significantly larger than that of remote sensing waterbody extraction datasets. If a large pretrained model is used in the RGB branch, it is unnecessary to freeze the weights. Instead, both branches can be trained jointly, allowing the RGB branch to better adapt to the remote sensing waterbody extraction task without raising concerns that fine-tuning might undermine the pretrained model’s strong representation capability. We also propose two X modality configurations with strong generalization performance. To fully fuse multimodal features, we design a hybrid fusion module combining a CNN and a cross-attention mechanism. To integrate the multi-scale features, we employ a multi-scale Transformer structure in the RGB branch and design a multi-scale decoder. Our algorithm achieves state-of-the-art performance on the GID-5 dataset and competitive performance on the S1S2-Water dataset. Furthermore, it significantly outperforms existing methods in cross-dataset zero-shot transfer between the two datasets, with IoU/F1-score gains of 26.08%/27.33% on GID-5 and 38.74%/31.37% on S1S2-Water over previous SOTA methods. Our processing paradigm of modeling RGB-NIR remote sensing images as RGB+X data shows potential for generalization to other multi-modal remote sensing tasks. The dual-complexity backbone we design also has potential to be extended to other tasks that transfer large pretrained RGB models to remote sensing imagery with RGB-NIR four bands or even more spectral bands. We have open-sourced the code and trained models used in this research. Full article

As climate change challenges intensify, the low-carbon transition has emerged as a fundamental structural transformation reshaping the global economic system and promoting sustainable development. In China, the “Dual Carbon” goals announced in September 2020 represent a landmark policy shift that imposes substantial environmental and regulatory pressure on high-carbon-emission enterprises. Against this backdrop, understanding how firms are adjusting their financial reporting practices to align with the low-carbon transition holds considerable significance for fostering their long-term sustainable development. Unlike previous studies that primarily attributed accounting conservatism to firm-specific risks or general economic uncertainty, this paper views the low-carbon transition as a structural institutional shock that reshapes firms’ external governance environment and information conditions, thereby offering a policy-driven explanation for accounting conservatism. Analysis using the Difference-in-differences method demonstrates that the low-carbon transition significantly enhances accounting conservatism among these enterprises (coefficient = 0.008, t = 4.13). Furthermore, mechanism analysis reveals that the low-carbon transition increases accounting conservatism through financing constraints and media attention. Heterogeneity analysis further indicates that the relationship between the low-carbon transition and accounting conservatism is more pronounced in non-state-owned enterprises, firms located in the eastern region, those facing intense industry competition, and companies with low levels of green innovation. Overall, the findings suggest that accounting conservatism is shaped not only by firm-level factors but also by large-scale institutional and policy transitions. By emphasizing that environmental regulation is a structural determinant of financial reporting behavior, this study extends the accounting conservatism literature. Furthermore, it demonstrates that improving financial reporting quality and risk identification capabilities enhances firms’ ability to address the challenges of the low-carbon transition, thereby fostering their long-term sustainable development. Full article

The key parameters of atmospheric convective stability, the K index (KI) and the Showalter index (SI), are important indicators for severe convective weather warnings. This study adopts a variational mode decomposition, bidirectional gated recurrent unit, and attention mechanism weighting combined model (VMD–BiGRU–Attention) to optimize core hyperparameters and verify model stability. Global Navigation Satellite System-derived precipitable water vapor (PWV) and relative humidity (RH) are incorporated as key parameters representing atmospheric water vapor conditions, thereby assisting VMD decomposition in accurately separating effective signals related to severe convection. The results show that the optimal VMD decomposition parameter K for the KI is 10 (minimum root mean square error [RMSE] = 3.96), whereas the optimal K for the SI is 11 (minimum RMSE = 1.87), verifying the applicability of VMD decomposition. In the validation using extreme rainfall events (2021–2025) at three meteorological stations in Guangxi (Baise, Nanning, and Hepu), the model, with the auxiliary contributions of PWV and RH, stably and accurately predicts the KI and SI for the next three hours, effectively capturing the critical characteristics of severe convection. The predicted results are consistent with the observed precipitation, demonstrating significant practical application value. Full article

Rock glaciers are important ice-debris landforms in high-mountain permafrost environments, but the development, knowledge base, and emerging directions of this research field remain insufficiently synthesized. This study retrieved English-language article and article/data paper records from the Science Citation Index Expanded database of the Web of Science Core Collection using the query TS = (“rock glacier*” OR “rock glacier*”). After document-type filtering and manual screening, 1125 valid records published between 1910 and 2025 were analyzed. Descriptive bibliometrics were used to characterize scientific production and collaboration patterns, Reference Publication Year Spectroscopy (RPYS) was used to identify historically influential publication years and foundational references, and keyword co-occurrence networks, thematic mapping, and thematic evolution analysis were used to trace associations among research topics. A Logistic life-cycle model was used only as a diagnostic tool for the current publication stage, not as a deterministic forecast. The results indicate that global rock glacier research remains in an active growth stage, although model-derived saturation values should be interpreted cautiously because bibliometric trajectories are affected by database coverage, indexing practices, research funding, technological change, and policy demand. RPYS shows that the knowledge base evolved from geomorphological description, classification, and genetic debate toward permafrost creep, internal structure, thermo-mechanical response, and hydrological significance. Keyword and thematic analyses show increasing attention to climate change, mountain permafrost, InSAR, ground-penetrating radar, hydrological processes, and multi-source monitoring. Because the dataset is restricted to English-language SCI-Expanded records, the results should be interpreted as a map of indexed international literature rather than a complete inventory of all rock glacier knowledge. Full article

Background/Objectives: Metabolic-dysfunction-associated steatotic liver disease (MASLD) and its progressive inflammatory/fibrotic form, metabolic-dysfunction-associated steatohepatitis (MASH), represent a growing global health burden. This progression is driven by complex mechanisms involving metabolic dysregulation, chronic inflammation, oxidative stress, and progressive fibrosis. To date, effective pharmacological therapies remain limited. Pentacyclic triterpenes have attracted increasing attention due to their broad biological activities and ability to modulate multiple molecular pathways implicated in chronic liver disease. This review aims to provide a mechanistic overview of the potential role of pentacyclic triterpenes in MASLD and MASH. Methods: A literature review was conducted using major scientific databases (PubMed and Web of Science) to identify experimental studies investigating pentacyclic triterpenes in metabolic liver diseases. Selected studies were analyzed according to triterpene structural classification, reported bioactivities, molecular targets, and experimental evidence from in vitro and in vivo models of MASLD/MASH or related pathogenic processes. Results: Pentacyclic triterpenes, especially ursolic acid, oleanolic acid, and glycyrrhizin, exhibit hepatoprotective effects including regulation of lipid metabolism, attenuation of oxidative and endoplasmic reticulum stress, suppression of pro-inflammatory signaling, inhibition of inflammasome activation, and reduction in hepatic stellate cell activation and extracellular matrix deposition. These effects involve modulation of signaling pathways, including AMPK, NF-κB, NLRP3, TGF-β, FXR, and MAPK. Preclinical evidence demonstrates improvements in steatosis, inflammation, and fibrosis in experimental models. Conclusions: Pentacyclic triterpenes emerge as multitarget modulators of MASH pathophysiology. However, translating preclinical evidence into well-designed clinical trials is necessary to validate their safety and efficacy in humans. Full article

Environmental pollution has emerged as one of the most significant threats to human and ecosystem health, with growing evidence suggesting that chronic exposure to toxic substances may accelerate aging. The concept of gerontogens, toxic compounds capable of accelerating this biological process, has gained increasing attention in toxicological research, particularly in the context of global demographic shifts toward older populations. Current research on gerontogens relies heavily on invertebrate models with short lifespans, such as Caenorhabditis elegans, Drosophila melanogaster, and Saccharomyces cerevisiae, which are valuable for studying conserved mechanisms in aging pathways, but present significant limitations for translational accuracy to many aspects of vertebrate biology. Vertebrate models traditionally employed in toxicology, including mice and zebrafish, require substantially longer experimental timelines and higher financial investments, making lifetime exposure and aging assays particularly challenging. In this context, the turquoise killifish Nothobranchius furzeri emerges as a highly promising vertebrate model for aging toxicology research. Recognized as the shortest-lived vertebrate species maintained under laboratory conditions, N. furzeri reaches sexual maturity within 14 days and displays complete senescence by 4 months of age, at which point individuals are considered elderly, offering a decisive advantage over conventional vertebrate models. Furthermore, its capacity for embryonic diapause enables practical embryo storage, long-distance transport, and synchronized hatching, greatly facilitating experimental designs. Although N. furzeri is well established in gerontological research, with studies addressing hallmarks of aging such as telomere shortening, neurodegeneration, and cellular senescence, its application in ecotoxicology remains remarkably limited, with fewer than 10 published studies to date. This article argues that N. furzeri may represent a critical bridge between toxicology and aging research, offering an efficient and translationally relevant platform for investigating the effects of environmental contaminants on vertebrate aging. Current limitations of the model, such as lack of husbandry standardization, are also discussed. Expanding its use in this field holds considerable potential for advancing evidence-based strategies in public health and environmental conservation related to chronic exposure to contaminants. Full article

With the rapid advancement of Earth observation technologies and the growing demand for intelligent remote sensing applications, high-resolution remote sensing imagery provides critical data support for a range of downstream applications, including land monitoring and disaster assessment. In this context, multi-label remote sensing image classification has become an important research task, because a single image may contain multiple ground-object categories with complex spatial distributions and semantic co-occurrence relationships. However, challenges such as the coexistence of multiscale objects, complex semantic dependencies, and long-tail category distributions impose significant limitations on existing methods in terms of feature representation capacity and class-balanced modeling. To address these challenges, a Multiscale Dynamic Reasoning Network (MSDR-Net) is proposed. Different from methods that focus on localized optimization for a single challenge, MSDR-Net establishes a task-driven modeling framework that jointly integrates multiscale feature extraction, label-aware semantic reasoning, and long-tail category optimization within an end-to-end architecture. The proposed network consists of three core modules. The Multiscale Feature Enhancement (MSFE) module incorporates a Feature Pyramid Network-based fusion mechanism, integrating deep semantic information with shallow, detailed features to effectively enhance the representation of multiscale objects. The Dynamic Semantic Reasoning (DSR) module introduces a Transformer-based global attention mechanism that models long-range dependencies among image features, enabling the capture of complex global semantic relationships. In the loss optimization stage, a Difficulty-Weighted Loss (DW-Loss) is introduced, which jointly incorporates category frequency weights and prior difficulty coefficients to dynamically regulate the contributions of rare classes and hard samples during training, thereby mitigating bias induced by class imbalance. Experiments conducted on the large-scale Detection in Optical Remote Sensing Images dataset demonstrate that the proposed method achieves superior performance. Ablation studies validate the effectiveness of each component, while comparative experiments indicate that MSDR-Net achieves a mean Average Precision of 95.88%, outperforming existing state-of-the-art methods. An improvement of approximately 1.74% is observed over the strongest baseline, MSCA, with consistent advantages demonstrated across Overall F1 and Class-wise F1 metrics. By unifying multiscale feature extraction, global semantic reasoning, and balanced loss optimization within a single framework, MSDR-Net provides a robust and efficient solution for multi-label classification in complex remote sensing scenarios. Full article

The segmentation of rocks on the Martian surface is crucial for navigation and obstacle avoidance by Mars rovers. However, frequent dust storms degrade rock surface textures, and the wide range of rock scales—from sub-meter to ten-meter—further complicates segmentation, especially under the strict computational constraints of rover hardware. This paper proposes a lightweight network named MAFT, specifically designed for Martian rock segmentation. The network builds upon the Adaptive Frequency Transformer (AFFormer) and constructs an improved backbone termed the Improved Adaptive Frequency Transformer (IAFFormer). By replacing the traditional self-attention mechanism with a frequency-domain approach, it captures global feature dependencies while reducing the computational complexity from quadratic to linear. The spatially isolated 1 × 1 convolutions in the pixel descriptor module are further replaced with Adaptive Kernel Convolution (AKConv), enabling the backbone to dynamically adjust its sampling positions to conform to the irregular and diverse morphologies of Martian rocks. An Enhanced Multidimensional Convolutional Attention (EMCA) module is introduced as the decoding structure. By integrating max-pooling in the squeeze stage and adaptive dilated convolutions in the excitation stage, EMCA strengthens the boundary perception and long-range dependency modeling of dust-covered rocks without increasing the parameter count. Additionally, we constructed a dataset of Martian rocks for the Zhurong rover (TWMARS-V2) and conducted experiments using a synthetic dataset (SynMars) and a real dataset (MarsData-V2). Experimental results demonstrate that MAFT achieves the highest segmentation accuracy among all compared methods, with only 2.97 M parameters and 15.49 G FLOPs. On the TWMARS-V2 dataset, Pixel Accuracy (PA) reaches 98.17%, and IoU reaches 88.90%. Full article

In practical imaging applications, low-light and overexposure are two common types of image degradation problems with inherent conflicts, and existing methods struggle to achieve accurate restoration of both degradations within a unified framework. To address this challenge, this paper proposes DFE-Net based on explicit frequency decoupling. The network adopts a symmetric U-Net architecture and embeds discrete wavelet transform (DWT) and inverse discrete wavelet transform (IWT) to construct an explicit dual-frequency processing mechanism, which optimizes the low-frequency information carrying global illumination and the high-frequency information containing detailed textures, respectively. In the encoder, DWT decouples features into low-frequency and high-frequency sub-bands and feeds them into dedicated enhancement modules. The low-frequency enhancement block integrates SS2D and a gated convolutional feed-forward network to efficiently model global contextual dependencies with linear complexity and accurately restore image illumination and contrast; the high-frequency enhancement block adopts CMT attention combined with a matching convolutional feed-forward network, enabling the detail restoration process to be guided by the optimized low-frequency information and ensuring the collaborative optimization of global structure and local textures. The decoder completes the reconstruction and fusion of the processed sub-bands through IWT. The quantitative and qualitative experimental results on the MSEC, SICE, and LOLv1 datasets demonstrate that DFE-Net achieves or surpasses existing state-of-the-art methods in various metrics while maintaining low model complexity. Full article

Accurate temperature forecasting is essential for agriculture, disaster management, and infrastructure planning, yet numerical weather prediction models face increasing limitations under climate change. Deep learning architectures, particularly transformers, offer promising alternatives but suffer from quadratic complexity and limited ability to capture both long-range dependencies and local periodic patterns in volatile climate data. To address these challenges, this paper proposes the Hybrid Attention Informer (HA-Informer), a unified end-to-end framework that introduces three modifications to the standard Informer: a hybrid attention mechanism combining ProbSparse attention with depthwise separable convolutions to capture global and local patterns simultaneously, an adaptive distillation mechanism that dynamically adjusts compression based on attention entropy to preserve fine-grained information, and a residual refinement decoder that mitigates error accumulation in long-horizon forecasting. The proposed model is evaluated on hourly temperature data from three climatically diverse cities, Niamey (hot), Tehran (temperate), and Harbin (cold), against seven baselines, namely, LSTM, a CNN, ARIMA, XGBoost, DLinear, transformer, and the standard Informer. The experimental results demonstrate that HA-Informer consistently achieves the lowest forecasting errors across all three locations, with mean squared error reductions of approximately $54\%$ over Informer, $85\%$ over DLinear, and $90\%$ over LSTM in the Niamey dataset, supported by statistically significant Diebold–Mariano test statistics ($p<0.05$) confirming the superiority of the proposed approach. Full article