The large-scale production of high-Ti steels is limited by the formation of Ti-containing oxides or nitrides in steel-slag reactions during continuous casting.These processes degrade mold flux properties,clog submerge...The large-scale production of high-Ti steels is limited by the formation of Ti-containing oxides or nitrides in steel-slag reactions during continuous casting.These processes degrade mold flux properties,clog submerged entry nozzles,form floaters in the molds,and produce various surface defects on the cast slabs.This review summarizes the effects of nonmetallic inclusions on traditional CaO-SiO_(2)-based(CS)mold fluxes and novel CaO-Al_(2)O_(3)-based(CA)low-or non-reactive fluxes containing TiO_(2),BaO,and B_(2)O_(3)additives to avoid undesirable steel,slag,and inclusion reactions,with the aim of providing a new perspective for research and practice related to balancing the lubrication and heat transfer of mold fluxes to promote smooth operation and reduce surface defects on cast slabs.For traditional CS mold flux,although the addition of solvents such as Na_(2)O,Li_(2)O,and B_(2)O_(3)can enhance flowability,steel-slag reactions persist,limiting the effectiveness of CS mold fluxes in high-Ti steel casting.Low-or non-reactive CA mold fluxes with reduced SiO_(2)content are a research focus,where adding other components can significantly change flux characteristics.Replacing CaO with BaO can lower the melting point and inhibit crystallization,allowing the flux to maintain good flowability at low temperatures.Replacing SiO_(2)with TiO_(2)can stabilize the viscosity and enhance heat transfer.To reduce the environmental impact,fluorides are replaced with components such as TiO_(2),B_(2)O_(3),BaO,Li_(2)O,and Na_(2)O for F-frce mold fluxes with similar lubrication,crystallization,and heat-transfer effects.When TiO_(2)replaces CaF_(2),it stabilizes the viscosity and enhances the heat conductivity,forming CaTiO_(3)and CaSiTiO_(5)phases instead of cuspidine to control crystallization.B_(2)O_(3)lowers the melting point and suppresses crystallization,forming phases such as Ca_(3)B_(2)O_(6)and Ca_(11)Si_(4)B_(2)O_(22).BaO introduces non-bridging oxygen to reduce viscosity and ensure flux flowability at low temperatures.However,further studies are required to determine the optimal mold flux compositions corresponding to the steel grades and the interactions between the various components of the mold flux.In the future,the practical application of new mold fluxes for high-Ti steel will become the focus of further verification to achieve a balance between lubrication and heat transfer,which is expected to minimize the occurrence of casting problems and slab defects.展开更多
Frequent droughts pose considerable threat to global forest carbon uptake,but little is known about the response of forest carbon fluxes in climatic transition zones to seasonal drought.In this study,the responses of ...Frequent droughts pose considerable threat to global forest carbon uptake,but little is known about the response of forest carbon fluxes in climatic transition zones to seasonal drought.In this study,the responses of carbon fluxes to seasonal drought in two natural forests(Quercus aliena var.acute serrata Maxim and Pinus tabuliformis Carr.)in the Baotianman Nature Reserve were investigated.The Q.aliena forest exhibited a high resilience with stable gross primary productivity(GPP).However,ecosystem respiration(Re)significantly declined by 18.4%compared with normal years,leading to an increase in net carbon sequestration capacity of 4.1%.This resilience was attributed to its deep root system accessing soil water(SWC_(50cm))to sustain stomatal openness,coupled with the efficient utilization of photosynthetically active radiation to drive photosynthesis.In contrast,the P.tabuliformis forest,which relied on shallow soil moisture(SWC_(20cm)),experienced simultaneous decreases in both GPP and Re during drought,with a sharply greater decrease in GPP,resulting in low net carbon sink capacity.Further analysis revealed that the Q.aliena forest prioritized carbon assimilation through a deep water-stomatal synergy strategy(anisohydric behavior),whereas the P.tabuliformis forest adopted an isohydric strategy favoring water conservation at the expense of carbon fixation efficiency.These findings highlight distinct mechanisms underlying drought adaptation between forest types,providing critical insight into optimizing forest carbon cycle models and selecting drought-resistant species under the influence of climate change.展开更多
Synchronous reluctance motors(SynRM)are widely employed in industrial applications due to their high robustness,low cost,and absence of permanent magnets.In recent years,significant research efforts have focused on im...Synchronous reluctance motors(SynRM)are widely employed in industrial applications due to their high robustness,low cost,and absence of permanent magnets.In recent years,significant research efforts have focused on improving the controllability and efficiency of SynRM.Accurate rotor position information is essential for the controller to generate appropriate current and voltage references corresponding to the desired speed and load torque.Shaft-mounted position sensors are generally undesirable because of their high cost,sensitivity to harsh operating conditions,maintenance requirements,and reduced reliability in environments characterized by high vibration.Consequently,sensorless control techniques that estimate rotor position using measured stator currents and voltages have attracted increasing attention.However,magnetic saturation,parameter nonlinearities,and cross-coupling effects significantly degrade position estimation accuracy and may compromise the stability of sensorless SynRM drives.In this paper,a nonlinear SynRM model is developed using finite element analysis(FEA)to accurately capture magnetic saturation and cross-coupling effects,thereby providing a precise representation of the machine’s electromagnetic behavior under varying load and flux conditions.A series of magnetostatic FEA simulations is performed.To reduce computational complexity,only one motor pole is analyzed by applying anti-periodic boundary conditions along the domain sides and enforcing a zero magnetic vector potential on the external stator boundary.Nonlinear iron material properties are modeled using the appropriate B-H curve.The simulations are carried out by imposing d-and q-axis current components and computing the corresponding flux linkages and electromagnetic torque.Based on these results,both apparent and incremental inductances are extracted and incorporated into the control algorithm.An advanced fictitious flux linkage method combined with a phase-locked loop(PLL)is employed for accurate rotor position estimation.Simulation results confirm that the proposed sensorless control strategy ensures stable operation and high position estimation accuracy over the entire speed range.展开更多
In this paper,we are concerned with the stability of traveling wavefronts of a Belousov-Zhabotinsky model with mixed nonlocal and degenerate diffusions.Such a system can be used to study the competition among nonlocal...In this paper,we are concerned with the stability of traveling wavefronts of a Belousov-Zhabotinsky model with mixed nonlocal and degenerate diffusions.Such a system can be used to study the competition among nonlocally diffusive species and degenerately diffusive species.We prove that the traveling wavefronts are exponentially stable,when the initial perturbation around the traveling waves decays exponentially as x→-∞,but in other locations,the initial data can be arbitrarily large.The adopted methods are the weighted energy with the comparison principle and squeezing technique.展开更多
With the rapid development of generative artificial intelligence(GenAI),the task of story visualization,which transforms natural language narratives into coherent and consistent image sequences,has attracted growing r...With the rapid development of generative artificial intelligence(GenAI),the task of story visualization,which transforms natural language narratives into coherent and consistent image sequences,has attracted growing research attention.However,existing methods still face limitations in balancing multi-frame character consistency and generation efficiency,which restricts their feasibility for large-scale practical applications.To address this issue,this study proposes a modular cloud-based distributed system built on Stable Diffusion.By separating the character generation and story generation processes,and integratingmulti-feature control techniques,a cachingmechanism,and an asynchronous task queue architecture,the system enhances generation efficiency and scalability.The experimental design includes both automated and human evaluations of character consistency,performance testing,and multinode simulation.The results show that the proposed system outperforms the baseline model StoryGen in both CLIP-I and human evaluation metrics.In terms of performance,under the experimental environment of this study,dual-node deployment reduces average waiting time by approximately 19%,while the four-node simulation further reduces it by up to 65%.Overall,this study demonstrates the advantages of cloud-distributed GenAI in maintaining character consistency and reducing generation latency,highlighting its potential value inmulti-user collaborative story visualization applications.展开更多
With the development of technology,diffusion model-based solvers have shown significant promise in solving Combinatorial Optimization(CO)problems,particularly in tackling Non-deterministic Polynomial-time hard(NP-hard...With the development of technology,diffusion model-based solvers have shown significant promise in solving Combinatorial Optimization(CO)problems,particularly in tackling Non-deterministic Polynomial-time hard(NP-hard)problems such as the Traveling Salesman Problem(TSP).However,existing diffusion model-based solvers typically employ a fixed,uniform noise schedule(e.g.,linear or cosine annealing)across all training instances,failing to fully account for the unique characteristics of each problem instance.To address this challenge,we present GraphGuided Diffusion Solvers(GGDS),an enhanced method for improving graph-based diffusion models.GGDS leverages Graph Neural Networks(GNNs)to capture graph structural information embedded in node coordinates and adjacency matrices,dynamically adjusting the noise levels in the diffusion model.This study investigates the TSP by examining two distinct time-step noise generation strategies:cosine annealing and a Neural Network(NN)-based approach.We evaluate their performance across different problem scales,particularly after integrating graph structural information.Experimental results indicate that GGDS outperforms previous methods with average performance improvements of 18.7%,6.3%,and 88.7%on TSP-500,TSP-100,and TSP-50,respectively.Specifically,GGDS demonstrates superior performance on TSP-500 and TSP-50,while its performance on TSP-100 is either comparable to or slightly better than that of previous methods,depending on the chosen noise schedule and decoding strategy.展开更多
Prussian blue analogs(PBAs)have emerged as environmentally friendly and structurally tunable cathode materials for aqueous ammonium-ion batteries(AIBs).However,the fundamental role of crystalline H_(2)O in regulating ...Prussian blue analogs(PBAs)have emerged as environmentally friendly and structurally tunable cathode materials for aqueous ammonium-ion batteries(AIBs).However,the fundamental role of crystalline H_(2)O in regulating ammonium-ion storage and transport remains poorly understood.In this study,we present a comprehensive comparison between hydrated NH_(4)NiHCF-H_(2)O and its anhydrous counterpart NH_(4)NiHCF,revealing the critical contribution of interstitial water to electrochemical performance.Structural and spectroscopic analyses confirm that interstitial water forms robust hydrogen bonds with NH_(4)+ions,stabilizing the PBA framework and mitigating structural degradation during cycling.Electrochemical measurements show that NH_(4)NiHCF-H_(2)O delivers a significantly higher specific capacity of 61 mA h g^(−1)at 0.2 C and markedly improved rate performance compared to NH_(4)NiHCF(48 mA h g^(−1)at 0.2 C).Kinetic analysis reveals that interstitial water enhances NH_(4)+diffusion,as evidenced by higher diffusion coefficients.Furthermore,density functional theory(DFT)calculations demonstrate that crystal water acts as a hydrogen bond acceptor,preferentially interacting with NH_(4)+and reducing the migration energy barrier,thereby facilitating fast ion transport.This work provides fundamental insights into the role of crystal water in PBAs and offers a rational design strategy for improving the kinetics,structural stability of PBAs cathodes for AIBs.展开更多
In this paper,we propose a new privacy-aware transmission scheduling algorithm for 6G ad hoc networks.This system enables end nodes to select the optimum time and scheme to transmit private data safely.In 6G dynamic h...In this paper,we propose a new privacy-aware transmission scheduling algorithm for 6G ad hoc networks.This system enables end nodes to select the optimum time and scheme to transmit private data safely.In 6G dynamic heterogeneous infrastructures,unstable links and non-uniform hardware capabilities create critical issues regarding security and privacy.Traditional protocols are often too computationally heavy to allow 6G services to achieve their expected Quality-of-Service(QoS).As the transport network is built of ad hoc nodes,there is no guarantee about their trustworthiness or behavior,and transversal functionalities are delegated to the extreme nodes.However,while security can be guaranteed in extreme-to-extreme solutions,privacy cannot,as all intermediate nodes still have to handle the data packets they are transporting.Besides,traditional schemes for private anonymous ad hoc communications are vulnerable against modern intelligent attacks based on learning models.The proposed scheme fulfills this gap.Findings show the probability of a successful intelligent attack reduces by up to 65%compared to ad hoc networks with no privacy protection strategy when used the proposed technology.While congestion probability can remain below 0.001%,as required in 6G services.展开更多
The neutron diffusion equation plays a pivotal role in nuclear reactor analysis.Nevertheless,employing the physics-informed neural network(PINN)method for its solution entails certain limitations.Conventional PINN app...The neutron diffusion equation plays a pivotal role in nuclear reactor analysis.Nevertheless,employing the physics-informed neural network(PINN)method for its solution entails certain limitations.Conventional PINN approaches generally utilize a fully connected network(FCN)architecture that is susceptible to overfitting,training instability,and gradient vanishing as the network depth increases.These challenges result in accuracy bottlenecks in the solution.In response to these issues,the residual-based resample physics-informed neural network(R2-PINN)is proposed.It is an improved PINN architecture that replaces the FCN with a convolutional neural network with a shortcut(S-CNN).It incorporates skip connections to facilitate gradient propagation between network layers.Additionally,the incorporation of the residual adaptive resampling(RAR)mechanism dynamically increases the number of sampling points.This,in turn,enhances the spatial representation capabilities and overall predictive accuracy of the model.The experimental results illustrate that our approach significantly improves the convergence capability of the model and achieves high-precision predictions of the physical fields.Compared with conventional FCN-based PINN methods,R 2-PINN effectively overcomes the limitations inherent in current methods.Thus,it provides more accurate and robust solutions for neutron diffusion equations.展开更多
Crack detection accuracy in computer vision is often constrained by limited annotated datasets.Although Generative Adversarial Networks(GANs)have been applied for data augmentation,they frequently introduce blurs and ...Crack detection accuracy in computer vision is often constrained by limited annotated datasets.Although Generative Adversarial Networks(GANs)have been applied for data augmentation,they frequently introduce blurs and artifacts.To address this challenge,this study leverages Denoising Diffusion Probabilistic Models(DDPMs)to generate high-quality synthetic crack images,enriching the training set with diverse and structurally consistent samples that enhance the crack segmentation.The proposed framework involves a two-stage pipeline:first,DDPMs are used to synthesize high-fidelity crack images that capture fine structural details.Second,these generated samples are combined with real data to train segmentation networks,thereby improving accuracy and robustness in crack detection.Compared with GAN-based approaches,DDPM achieved the best fidelity,with the highest Structural Similarity Index(SSIM)(0.302)and lowest Learned Perceptual Image Patch Similarity(LPIPS)(0.461),producing artifact-free images that preserve fine crack details.To validate its effectiveness,six segmentation models were tested,among which LinkNet consistently achieved the best performance,excelling in both region-level accuracy and structural continuity.Incorporating DDPM-augmented data further enhanced segmentation outcomes,increasing F1 scores by up to 1.1%and IoU by 1.7%,while also improving boundary alignment and skeleton continuity compared with models trained on real images alone.Experiments with varying augmentation ratios showed consistent improvements,with F1 rising from 0.946(no augmentation)to 0.957 and IoU from 0.897 to 0.913 at the highest ratio.These findings demonstrate the effectiveness of diffusion-based augmentation for complex crack detection in structural health monitoring.展开更多
The collection and annotation of lar ge-scale bird datasets are resource-intensive and time-consuming processes that significantly limit the scalability and accuracy of biodiversity monitoring systems.While self-super...The collection and annotation of lar ge-scale bird datasets are resource-intensive and time-consuming processes that significantly limit the scalability and accuracy of biodiversity monitoring systems.While self-supervised learning(SSL)has emerged as a promising approach for leveraging unannotated data,current SSL methods face two critical challenges in bird species recognition:(1)long-tailed data distributions that result in poor performance on underrepresented species;and(2)domain shift issues caused by data augmentation strategies designed to mitigate class imbalance.Here we present SDNet,a novel SSL-based bird recognition framework that integrates diffusion models with large language models(LLMs)to overcome these limitations.SDNet employs LLMs to generate semantically rich textual descriptions for tail-class species by prompting the models with species taxonomy,morphological attributes,and habitat information,producing detailed natural language priors that capture fine-grained visual characteristics(e.g.,plumage patterns,body proportions,and distinctive markings).These textual descriptions are subsequently used by a conditional diffusion model to synthesize new bird image samples through cross-attention mechanisms that fuse textual embeddings with intermediate visual feature representations during the denoising process,ensuring generated images preserve species-specific morphological details while maintaining photorealistic quality.Additionally,we incorporate a Swin Transformer as the feature extraction backbone whose hierarchical window-based attention mechanism and shifted windowing scheme enable multi-scale local feature extraction that proves particularly effective at capturing finegrained discriminative patterns(such as beak shape and feather texture)while mitigating domain shift between synthetic and original images through consistent feature representations across both data sources.SDNet is validated on both a self-constructed dataset(Bird_BXS)an d a publicly available benchmark(Birds_25),demonstrating substantial improvements over conventional SSL approaches.Our results indicate that the synergistic integration of LLMs,diffusion models,and the Swin Transformer architecture contributes significantly to recognition accuracy,particularly for rare and morphologically similar species.These findings highlight the potential of SDNet for addressing fundamental limitations of existing SSL methods in avian recognition tasks and establishing a new paradigm for efficient self-supervised learning in large-scale ornithological vision applications.展开更多
Lacustrine groundwater discharge(LGD)plays an important role in water resources management.Previous studies have focused on LGD process in a single lake,but the differences in LGD process within the same region have n...Lacustrine groundwater discharge(LGD)plays an important role in water resources management.Previous studies have focused on LGD process in a single lake,but the differences in LGD process within the same region have not been thoroughly investigated.In this study,multiple tracers(hydrochemistry,𝛿D,𝛿18O and 222Rn)were used to compare mechanisms of LGD in Daihai and Ulansuhai Lake in Inner Mongoli1,Northwest China.The hydrochemical types showed a trend from groundwater to lake water,indicating a hydraulic connection between them.In addition,the𝛿D and𝛿18O values of sediment pore water were between the groundwater and lake water,indicating the LGD processes.The radon mass balance model was used to estimate the average groundwater discharge rates of Daihai and Ulansuhai Lake,which were 2.79 mm/day and 3.02 mm/day,respectively.The total nitrogen(TN),total phosphorus(TP),and fluoride inputs associated with LGD in Daihai Lake accounted for 97.52%,96.59%,and 95.84%of the total inputs,respectively.In contrast,TN,TP and fluoride inputs in Ulansuhai Lake were 53.56%,40.98%,and 36.25%,respectively.This indicates that the pollutant inputs associated with LGD posed a potential threat to the ecological stability of Daihai and Ulansuhai Lake.By comparison,the differences of LGD process and associated pollutant flux were controlled by hydrogeological conditions,lakebed permeability and human activities.This study provides a reference for water resources management in Daihai and Ulansuhai Lake basins while improving the understanding of LGD in the Yellow River basin.展开更多
Amidst the growing global emphasis on nuclear safety,the integrity of nuclear reactor systems has garnered attention in the aftermath of consequential events.Moreover,the rapid development of artificial intelligence t...Amidst the growing global emphasis on nuclear safety,the integrity of nuclear reactor systems has garnered attention in the aftermath of consequential events.Moreover,the rapid development of artificial intelligence technology has provided immense opportunities to enhance the safety and economy of nuclear energy.However,data-driven deep learning techniques often lack interpretability,which hinders their applicability in the nuclear energy sector.To address this problem,this study proposes a hybrid data-driven and knowledge-driven artificial intelligence model based on physics-informed neural networks to accurately compute the neutron flux distribution inside a nuclear reactor core.Innovative techniques,such as regional decomposition,intelligent k_(eff)(effective multiplication factor)search,and k_(eff)inversion,have been introduced for the calculation.Furthermore,hyperparameters of the model are automatically optimized using a whale optimization algorithm.A series of computational examples are used to validate the proposed model,demonstrating its applicability,generality,and high accuracy in calculating the neutron flux within the nuclear reactor.The model offers a dependable strategy for computing the neutron flux distribution in nuclear reactors for advanced simulation techniques in the future,including reactor digital twinning.This approach is data-light,requires little to no training data,and still delivers remarkably precise output data.展开更多
Background:Medical imaging advancements are constrained by fundamental trade-offs between acquisition speed,radiation dose,and image quality,forcing clinicians to work with noisy,incomplete data.Existing reconstructio...Background:Medical imaging advancements are constrained by fundamental trade-offs between acquisition speed,radiation dose,and image quality,forcing clinicians to work with noisy,incomplete data.Existing reconstruction methods either compromise on accuracy with iterative algorithms or suffer from limited generalizability with task-specific deep learning approaches.Methods:We present LDM-PIR,a lightweight physics-conditioned diffusion multi-model for medical image reconstruction that addresses key challenges in magnetic resonance imaging(MRI),CT,and low-photon imaging.Unlike traditional iterative methods,which are computationally expensive,or task-specific deep learning approaches lacking generalizability,integrates three innovations.A physics-conditioned diffusion framework that embeds acquisition operators(Fourier/Radon transforms)and noise models directly into the reconstruction process.A multi-model architecture that unifies denoising,inpainting,and super-resolution via shared weight conditioning.A lightweight design(2.1M parameters)enabling rapid inference(0.8s/image on GPU).Through self-supervised fine-tuning with measurement consistency losses adapts to new imaging modalities using fewer annotated samples.Results:Achieves state-of-the-art performance on fastMRI(peak signal-to-noise ratio(PSNR):34.04 for single-coil/31.50 for multi-coil)and Lung Image Database Consortium and Image Database Resource Initiative(28.83 PSNR under Poisson noise).Clinical evaluations demonstrate superior preservation of anatomical structures,with SSIM improvements of 8.8%for single-coil and 4.36%for multi-coil MRI over uDPIR.Conclusion:It offers a flexible,efficient,and scalable solution for medical image reconstruction,addressing the challenges of noise,undersampling,and modality generalization.The model’s lightweight design allows for rapid inference,while its self-supervised fine-tuning capability minimizes reliance on large annotated datasets,making it suitable for real-world clinical applications.展开更多
The present study investigates the flow,heat,and mass transfer analysis in the bioconvection of nanofluid containing motile gyrotactic microorganisms through a semi-porous curved oscillatory channel with a magnetic fi...The present study investigates the flow,heat,and mass transfer analysis in the bioconvection of nanofluid containing motile gyrotactic microorganisms through a semi-porous curved oscillatory channel with a magnetic field.These microorganisms produce density gradients by swimming,which induces macroscopic convection flows in the fluid.This procedure improves the mass and heat transfer,illustrating the interaction between biological activity and fluid dynamics.Furthermore,instead of considering traditional Fourier's and Fick's law the energy and concentration equations are developed by incorporating Cattaneo-Christov double diffusion theory.Moreover,to examine the influence of thermophoresis and Brownian diffusions in the fluid we have adopted the Buongiorno nanofluid model.Due to the oscillation of the surface of the channel,the mathematical development of the considered flow problem is obtained in the form of partial differential equations via the curvilinear coordinate system.The convergent series solution of the governing flow equations is obtained after applying the homotopy analysis method(HAM).The effects of different pertinent flow parameters on velocity,motile microorganism density distribution,concentration,pressure,temperature,and skin friction coefficient are examined and discussed in detail with the help of graphs and tables.It is observed during the current study that the density of microorganisms is enhanced for higher values of Reynolds number,Peclet number,radius of curvature variable,and Lewis number.展开更多
Human motion modeling is a core technology in computer animation,game development,and humancomputer interaction.In particular,generating natural and coherent in-between motion using only the initial and terminal frame...Human motion modeling is a core technology in computer animation,game development,and humancomputer interaction.In particular,generating natural and coherent in-between motion using only the initial and terminal frames remains a fundamental yet unresolved challenge.Existing methods typically rely on dense keyframe inputs or complex prior structures,making it difficult to balance motion quality and plausibility under conditions such as sparse constraints,long-term dependencies,and diverse motion styles.To address this,we propose a motion generation framework based on a frequency-domain diffusion model,which aims to better model complex motion distributions and enhance generation stability under sparse conditions.Our method maps motion sequences to the frequency domain via the Discrete Cosine Transform(DCT),enabling more effective modeling of low-frequency motion structures while suppressing high-frequency noise.A denoising network based on self-attention is introduced to capture long-range temporal dependencies and improve global structural awareness.Additionally,a multi-objective loss function is employed to jointly optimize motion smoothness,pose diversity,and anatomical consistency,enhancing the realism and physical plausibility of the generated sequences.Comparative experiments on the Human3.6M and LaFAN1 datasets demonstrate that our method outperforms state-of-the-art approaches across multiple performance metrics,showing stronger capabilities in generating intermediate motion frames.This research offers a new perspective and methodology for human motion generation and holds promise for applications in character animation,game development,and virtual interaction.展开更多
Physics-informed neural networks(PINNs),as a novel artificial intelligence method for solving partial differential equations,are applicable to solve both forward and inverse problems.This study evaluates the performan...Physics-informed neural networks(PINNs),as a novel artificial intelligence method for solving partial differential equations,are applicable to solve both forward and inverse problems.This study evaluates the performance of PINNs in solving the temperature diffusion equation of the seawater across six scenarios,including forward and inverse problems under three different boundary conditions.Results demonstrate that PINNs achieved consistently higher accuracy with the Dirichlet and Neumann boundary conditions compared to the Robin boundary condition for both forward and inverse problems.Inaccurate weighting of terms in the loss function can reduce model accuracy.Additionally,the sensitivity of model performance to the positioning of sampling points varied between different boundary conditions.In particular,the model under the Dirichlet boundary condition exhibited superior robustness to variations in point positions during the solutions of inverse problems.In contrast,for the Neumann and Robin boundary conditions,accuracy declines when points were sampled from identical positions or at the same time.Subsequently,the Argo observations were used to reconstruct the vertical diffusion of seawater temperature in the north-central Pacific for the applicability of PINNs in the real ocean.The PINNs successfully captured the vertical diffusion characteristics of seawater temperature,reflected the seasonal changes of vertical temperature under different topographic conditions,and revealed the influence of topography on the temperature diffusion coefficient.The PINNs were proved effective in solving the temperature diffusion equation of seawater with limited data,providing a promising technique for simulating or predicting ocean phenomena using sparse observations.展开更多
Some patients with systemic lupus erythematosus experience neuropsychiatric symptoms.Although magnetic resonance imaging can detect abnormal signals in the white matter of the brain,conventional methods often struggle...Some patients with systemic lupus erythematosus experience neuropsychiatric symptoms.Although magnetic resonance imaging can detect abnormal signals in the white matter of the brain,conventional methods often struggle to accurately capture microstructural changes.Various diffusion models have been used to study white matter in systemic lupus erythematosus;however,comparative analyses of their sensitivity and specificity for detecting microstructural changes remain insufficient.To address this,our team designed a diagnostic trial that used multimodal diffusion imaging techniques to observe white matter microstructural changes in patients with systemic lupus erythematosus who had neuropsychiatric symptoms,with an aim to identify key diagnostic biomarkers for these patients.Patients with active lupus who received treatment at the Department of Rheumatology and Immunology,The First Affiliated Hospital of China Medical University,from September 2023 to March 2024 were recruited.According to the standards of the American College of Rheumatology,patients with systemic lupus erythematosus who had neuropsychiatric symptoms were assigned to the systemic lupus erythematosus group,whereas those without neuropsychiatric symptoms were assigned to the non-systemic lupus erythematosus group.Additionally,healthy volunteers matched by region,sex,and age were recruited as controls.All three groups underwent the same diffusion magnetic resonance imaging examination protocol to compare differences in diffusion parameters.Advanced diffusion imaging models were able to sensitively detect microstructural changes in the white matter fibers of patients with systemic lupus erythematosus who had neuropsychiatric symptoms,with specific diffusion parameters showing significant abnormalities in key brain regions.In the left superior longitudinal fasciculus subregion and the right thalamic radiations of patients with systemic lupus erythematosus who had neuropsychiatric symptoms,we also identified abnormal diffusion characteristics that were clearly correlated with disease activity,suggesting that microstructural changes in these areas may reflect the dynamic process of neuroinflammatory damage.The present study addresses critical challenges in the diagnosis of systemic lupus erythematosus by identifying specific white matter imaging biomarkers and elucidating the association between microstructural damage and clinical manifestations.The main contributions of our study include:1)establishing axial regression probability parameters from mean apparent propagator magnetic resonance imaging as sensitive biomarkers for systemic lupus erythematosus,particularly in the third subregion of the left superior longitudinal fasciculus;2)demonstrating that multimodal diffusion imaging may be superior to conventional diffusion tensor imaging for detecting white matter microstructural abnormalities in patients with systemic lupus erythematosus;and 3)integrating tract-based spatial statistics with clinically relevant analyses to link imaging findings to pathological mechanisms.展开更多
The internal flow fields within a three-dimensional inward-tunning combined inlet are extremely complex,especially during the engine mode transition,where the tunnel changes may impact the flow fields significantly.To...The internal flow fields within a three-dimensional inward-tunning combined inlet are extremely complex,especially during the engine mode transition,where the tunnel changes may impact the flow fields significantly.To develop an efficient flow field reconstruction model for this,we present an Improved Conditional Denoising Diffusion Generative Adversarial Network(ICDDGAN),which integrates Conditional Denoising Diffusion Probabilistic Models(CDDPMs)with Style GAN,and introduce a reconstruction discrimination mechanism and dynamic loss weight learning strategy.We establish the Mach number flow field dataset by numerical simulation at various backpressures for the mode transition process from turbine mode to ejector ramjet mode at Mach number 2.5.The proposed ICDDGAN model,given only sparse parameter information,can rapidly generate high-quality Mach number flow fields without a large number of samples for training.The results show that ICDDGAN is superior to CDDGAN in terms of training convergence and stability.Moreover,the interpolation and extrapolation test results during backpressure conditions show that ICDDGAN can accurately and quickly reconstruct Mach number fields at various tunnel slice shapes,with a Structural Similarity Index Measure(SSIM)of over 0.96 and a Mean-Square Error(MSE)of 0.035%to actual flow fields,reducing time costs by 7-8 orders of magnitude compared to Computational Fluid Dynamics(CFD)calculations.This can provide an efficient means for rapid computation of complex flow fields.展开更多
基金financially supported by the National Natural Science Foundation of China(Nos.52204345 and 52474361)the Scientific Research Innovation Projects of Graduate Student of Jiangsu province,China(No.KYCX24_4184)。
文摘The large-scale production of high-Ti steels is limited by the formation of Ti-containing oxides or nitrides in steel-slag reactions during continuous casting.These processes degrade mold flux properties,clog submerged entry nozzles,form floaters in the molds,and produce various surface defects on the cast slabs.This review summarizes the effects of nonmetallic inclusions on traditional CaO-SiO_(2)-based(CS)mold fluxes and novel CaO-Al_(2)O_(3)-based(CA)low-or non-reactive fluxes containing TiO_(2),BaO,and B_(2)O_(3)additives to avoid undesirable steel,slag,and inclusion reactions,with the aim of providing a new perspective for research and practice related to balancing the lubrication and heat transfer of mold fluxes to promote smooth operation and reduce surface defects on cast slabs.For traditional CS mold flux,although the addition of solvents such as Na_(2)O,Li_(2)O,and B_(2)O_(3)can enhance flowability,steel-slag reactions persist,limiting the effectiveness of CS mold fluxes in high-Ti steel casting.Low-or non-reactive CA mold fluxes with reduced SiO_(2)content are a research focus,where adding other components can significantly change flux characteristics.Replacing CaO with BaO can lower the melting point and inhibit crystallization,allowing the flux to maintain good flowability at low temperatures.Replacing SiO_(2)with TiO_(2)can stabilize the viscosity and enhance heat transfer.To reduce the environmental impact,fluorides are replaced with components such as TiO_(2),B_(2)O_(3),BaO,Li_(2)O,and Na_(2)O for F-frce mold fluxes with similar lubrication,crystallization,and heat-transfer effects.When TiO_(2)replaces CaF_(2),it stabilizes the viscosity and enhances the heat conductivity,forming CaTiO_(3)and CaSiTiO_(5)phases instead of cuspidine to control crystallization.B_(2)O_(3)lowers the melting point and suppresses crystallization,forming phases such as Ca_(3)B_(2)O_(6)and Ca_(11)Si_(4)B_(2)O_(22).BaO introduces non-bridging oxygen to reduce viscosity and ensure flux flowability at low temperatures.However,further studies are required to determine the optimal mold flux compositions corresponding to the steel grades and the interactions between the various components of the mold flux.In the future,the practical application of new mold fluxes for high-Ti steel will become the focus of further verification to achieve a balance between lubrication and heat transfer,which is expected to minimize the occurrence of casting problems and slab defects.
基金financially supported by the National Key Research and Development Program of China(2021YFD2200405)the National Natural Science Foundation of China(31930078)special funds for Baotianman Forest Ecosystem Research Station from Chinese Academy of Forestry and Ministry of Science and Technology of China。
文摘Frequent droughts pose considerable threat to global forest carbon uptake,but little is known about the response of forest carbon fluxes in climatic transition zones to seasonal drought.In this study,the responses of carbon fluxes to seasonal drought in two natural forests(Quercus aliena var.acute serrata Maxim and Pinus tabuliformis Carr.)in the Baotianman Nature Reserve were investigated.The Q.aliena forest exhibited a high resilience with stable gross primary productivity(GPP).However,ecosystem respiration(Re)significantly declined by 18.4%compared with normal years,leading to an increase in net carbon sequestration capacity of 4.1%.This resilience was attributed to its deep root system accessing soil water(SWC_(50cm))to sustain stomatal openness,coupled with the efficient utilization of photosynthetically active radiation to drive photosynthesis.In contrast,the P.tabuliformis forest,which relied on shallow soil moisture(SWC_(20cm)),experienced simultaneous decreases in both GPP and Re during drought,with a sharply greater decrease in GPP,resulting in low net carbon sink capacity.Further analysis revealed that the Q.aliena forest prioritized carbon assimilation through a deep water-stomatal synergy strategy(anisohydric behavior),whereas the P.tabuliformis forest adopted an isohydric strategy favoring water conservation at the expense of carbon fixation efficiency.These findings highlight distinct mechanisms underlying drought adaptation between forest types,providing critical insight into optimizing forest carbon cycle models and selecting drought-resistant species under the influence of climate change.
文摘Synchronous reluctance motors(SynRM)are widely employed in industrial applications due to their high robustness,low cost,and absence of permanent magnets.In recent years,significant research efforts have focused on improving the controllability and efficiency of SynRM.Accurate rotor position information is essential for the controller to generate appropriate current and voltage references corresponding to the desired speed and load torque.Shaft-mounted position sensors are generally undesirable because of their high cost,sensitivity to harsh operating conditions,maintenance requirements,and reduced reliability in environments characterized by high vibration.Consequently,sensorless control techniques that estimate rotor position using measured stator currents and voltages have attracted increasing attention.However,magnetic saturation,parameter nonlinearities,and cross-coupling effects significantly degrade position estimation accuracy and may compromise the stability of sensorless SynRM drives.In this paper,a nonlinear SynRM model is developed using finite element analysis(FEA)to accurately capture magnetic saturation and cross-coupling effects,thereby providing a precise representation of the machine’s electromagnetic behavior under varying load and flux conditions.A series of magnetostatic FEA simulations is performed.To reduce computational complexity,only one motor pole is analyzed by applying anti-periodic boundary conditions along the domain sides and enforcing a zero magnetic vector potential on the external stator boundary.Nonlinear iron material properties are modeled using the appropriate B-H curve.The simulations are carried out by imposing d-and q-axis current components and computing the corresponding flux linkages and electromagnetic torque.Based on these results,both apparent and incremental inductances are extracted and incorporated into the control algorithm.An advanced fictitious flux linkage method combined with a phase-locked loop(PLL)is employed for accurate rotor position estimation.Simulation results confirm that the proposed sensorless control strategy ensures stable operation and high position estimation accuracy over the entire speed range.
基金Supported by the National Natural Science Foundation of China(Grant No.12261081).
文摘In this paper,we are concerned with the stability of traveling wavefronts of a Belousov-Zhabotinsky model with mixed nonlocal and degenerate diffusions.Such a system can be used to study the competition among nonlocally diffusive species and degenerately diffusive species.We prove that the traveling wavefronts are exponentially stable,when the initial perturbation around the traveling waves decays exponentially as x→-∞,but in other locations,the initial data can be arbitrarily large.The adopted methods are the weighted energy with the comparison principle and squeezing technique.
文摘With the rapid development of generative artificial intelligence(GenAI),the task of story visualization,which transforms natural language narratives into coherent and consistent image sequences,has attracted growing research attention.However,existing methods still face limitations in balancing multi-frame character consistency and generation efficiency,which restricts their feasibility for large-scale practical applications.To address this issue,this study proposes a modular cloud-based distributed system built on Stable Diffusion.By separating the character generation and story generation processes,and integratingmulti-feature control techniques,a cachingmechanism,and an asynchronous task queue architecture,the system enhances generation efficiency and scalability.The experimental design includes both automated and human evaluations of character consistency,performance testing,and multinode simulation.The results show that the proposed system outperforms the baseline model StoryGen in both CLIP-I and human evaluation metrics.In terms of performance,under the experimental environment of this study,dual-node deployment reduces average waiting time by approximately 19%,while the four-node simulation further reduces it by up to 65%.Overall,this study demonstrates the advantages of cloud-distributed GenAI in maintaining character consistency and reducing generation latency,highlighting its potential value inmulti-user collaborative story visualization applications.
基金supported by the National Science and Technology Council,Taiwan,under grant no.NSTC 114-2221-E-197-005-MY3.
文摘With the development of technology,diffusion model-based solvers have shown significant promise in solving Combinatorial Optimization(CO)problems,particularly in tackling Non-deterministic Polynomial-time hard(NP-hard)problems such as the Traveling Salesman Problem(TSP).However,existing diffusion model-based solvers typically employ a fixed,uniform noise schedule(e.g.,linear or cosine annealing)across all training instances,failing to fully account for the unique characteristics of each problem instance.To address this challenge,we present GraphGuided Diffusion Solvers(GGDS),an enhanced method for improving graph-based diffusion models.GGDS leverages Graph Neural Networks(GNNs)to capture graph structural information embedded in node coordinates and adjacency matrices,dynamically adjusting the noise levels in the diffusion model.This study investigates the TSP by examining two distinct time-step noise generation strategies:cosine annealing and a Neural Network(NN)-based approach.We evaluate their performance across different problem scales,particularly after integrating graph structural information.Experimental results indicate that GGDS outperforms previous methods with average performance improvements of 18.7%,6.3%,and 88.7%on TSP-500,TSP-100,and TSP-50,respectively.Specifically,GGDS demonstrates superior performance on TSP-500 and TSP-50,while its performance on TSP-100 is either comparable to or slightly better than that of previous methods,depending on the chosen noise schedule and decoding strategy.
基金supported by the National Natural Science Foundation of China (52172227)the Natural Science Foundation of Hubei Province (2023AFA114)+2 种基金the Guizhou Provincial Key Technology R&D Program (ZD[2025]019)provided by the Startup Fund (20QD80 and 22QD28)support from the Science&Technology Top Talents Program of Guizhou Province ([2024]349)
文摘Prussian blue analogs(PBAs)have emerged as environmentally friendly and structurally tunable cathode materials for aqueous ammonium-ion batteries(AIBs).However,the fundamental role of crystalline H_(2)O in regulating ammonium-ion storage and transport remains poorly understood.In this study,we present a comprehensive comparison between hydrated NH_(4)NiHCF-H_(2)O and its anhydrous counterpart NH_(4)NiHCF,revealing the critical contribution of interstitial water to electrochemical performance.Structural and spectroscopic analyses confirm that interstitial water forms robust hydrogen bonds with NH_(4)+ions,stabilizing the PBA framework and mitigating structural degradation during cycling.Electrochemical measurements show that NH_(4)NiHCF-H_(2)O delivers a significantly higher specific capacity of 61 mA h g^(−1)at 0.2 C and markedly improved rate performance compared to NH_(4)NiHCF(48 mA h g^(−1)at 0.2 C).Kinetic analysis reveals that interstitial water enhances NH_(4)+diffusion,as evidenced by higher diffusion coefficients.Furthermore,density functional theory(DFT)calculations demonstrate that crystal water acts as a hydrogen bond acceptor,preferentially interacting with NH_(4)+and reducing the migration energy barrier,thereby facilitating fast ion transport.This work provides fundamental insights into the role of crystal water in PBAs and offers a rational design strategy for improving the kinetics,structural stability of PBAs cathodes for AIBs.
基金funding from the European Commission by the Ruralities project(grant agreement no.101060876).
文摘In this paper,we propose a new privacy-aware transmission scheduling algorithm for 6G ad hoc networks.This system enables end nodes to select the optimum time and scheme to transmit private data safely.In 6G dynamic heterogeneous infrastructures,unstable links and non-uniform hardware capabilities create critical issues regarding security and privacy.Traditional protocols are often too computationally heavy to allow 6G services to achieve their expected Quality-of-Service(QoS).As the transport network is built of ad hoc nodes,there is no guarantee about their trustworthiness or behavior,and transversal functionalities are delegated to the extreme nodes.However,while security can be guaranteed in extreme-to-extreme solutions,privacy cannot,as all intermediate nodes still have to handle the data packets they are transporting.Besides,traditional schemes for private anonymous ad hoc communications are vulnerable against modern intelligent attacks based on learning models.The proposed scheme fulfills this gap.Findings show the probability of a successful intelligent attack reduces by up to 65%compared to ad hoc networks with no privacy protection strategy when used the proposed technology.While congestion probability can remain below 0.001%,as required in 6G services.
基金supported by the Science and Technology on Reactor System Design Technology Laboratory(No.LRSDT12023108)supported in part by the Chongqing Postdoctoral Science Foundation(No.cstc2021jcyj-bsh0252)+2 种基金the National Natural Science Foundation of China(No.12005030)Sichuan Province to unveil the list of marshal industry common technology research projects(No.23jBGOV0001)Special Program for Stabilizing Support to Basic Research of National Basic Research Institutes(No.WDZC-2023-05-03-05).
文摘The neutron diffusion equation plays a pivotal role in nuclear reactor analysis.Nevertheless,employing the physics-informed neural network(PINN)method for its solution entails certain limitations.Conventional PINN approaches generally utilize a fully connected network(FCN)architecture that is susceptible to overfitting,training instability,and gradient vanishing as the network depth increases.These challenges result in accuracy bottlenecks in the solution.In response to these issues,the residual-based resample physics-informed neural network(R2-PINN)is proposed.It is an improved PINN architecture that replaces the FCN with a convolutional neural network with a shortcut(S-CNN).It incorporates skip connections to facilitate gradient propagation between network layers.Additionally,the incorporation of the residual adaptive resampling(RAR)mechanism dynamically increases the number of sampling points.This,in turn,enhances the spatial representation capabilities and overall predictive accuracy of the model.The experimental results illustrate that our approach significantly improves the convergence capability of the model and achieves high-precision predictions of the physical fields.Compared with conventional FCN-based PINN methods,R 2-PINN effectively overcomes the limitations inherent in current methods.Thus,it provides more accurate and robust solutions for neutron diffusion equations.
基金the National Natural Science Foundation of China(Grant No.:52508343)the Fundamental Research Funds for the Central Universities(Grant No.:B250201004).
文摘Crack detection accuracy in computer vision is often constrained by limited annotated datasets.Although Generative Adversarial Networks(GANs)have been applied for data augmentation,they frequently introduce blurs and artifacts.To address this challenge,this study leverages Denoising Diffusion Probabilistic Models(DDPMs)to generate high-quality synthetic crack images,enriching the training set with diverse and structurally consistent samples that enhance the crack segmentation.The proposed framework involves a two-stage pipeline:first,DDPMs are used to synthesize high-fidelity crack images that capture fine structural details.Second,these generated samples are combined with real data to train segmentation networks,thereby improving accuracy and robustness in crack detection.Compared with GAN-based approaches,DDPM achieved the best fidelity,with the highest Structural Similarity Index(SSIM)(0.302)and lowest Learned Perceptual Image Patch Similarity(LPIPS)(0.461),producing artifact-free images that preserve fine crack details.To validate its effectiveness,six segmentation models were tested,among which LinkNet consistently achieved the best performance,excelling in both region-level accuracy and structural continuity.Incorporating DDPM-augmented data further enhanced segmentation outcomes,increasing F1 scores by up to 1.1%and IoU by 1.7%,while also improving boundary alignment and skeleton continuity compared with models trained on real images alone.Experiments with varying augmentation ratios showed consistent improvements,with F1 rising from 0.946(no augmentation)to 0.957 and IoU from 0.897 to 0.913 at the highest ratio.These findings demonstrate the effectiveness of diffusion-based augmentation for complex crack detection in structural health monitoring.
基金supported by the National Natural Science Foundation of China(32471964)。
文摘The collection and annotation of lar ge-scale bird datasets are resource-intensive and time-consuming processes that significantly limit the scalability and accuracy of biodiversity monitoring systems.While self-supervised learning(SSL)has emerged as a promising approach for leveraging unannotated data,current SSL methods face two critical challenges in bird species recognition:(1)long-tailed data distributions that result in poor performance on underrepresented species;and(2)domain shift issues caused by data augmentation strategies designed to mitigate class imbalance.Here we present SDNet,a novel SSL-based bird recognition framework that integrates diffusion models with large language models(LLMs)to overcome these limitations.SDNet employs LLMs to generate semantically rich textual descriptions for tail-class species by prompting the models with species taxonomy,morphological attributes,and habitat information,producing detailed natural language priors that capture fine-grained visual characteristics(e.g.,plumage patterns,body proportions,and distinctive markings).These textual descriptions are subsequently used by a conditional diffusion model to synthesize new bird image samples through cross-attention mechanisms that fuse textual embeddings with intermediate visual feature representations during the denoising process,ensuring generated images preserve species-specific morphological details while maintaining photorealistic quality.Additionally,we incorporate a Swin Transformer as the feature extraction backbone whose hierarchical window-based attention mechanism and shifted windowing scheme enable multi-scale local feature extraction that proves particularly effective at capturing finegrained discriminative patterns(such as beak shape and feather texture)while mitigating domain shift between synthetic and original images through consistent feature representations across both data sources.SDNet is validated on both a self-constructed dataset(Bird_BXS)an d a publicly available benchmark(Birds_25),demonstrating substantial improvements over conventional SSL approaches.Our results indicate that the synergistic integration of LLMs,diffusion models,and the Swin Transformer architecture contributes significantly to recognition accuracy,particularly for rare and morphologically similar species.These findings highlight the potential of SDNet for addressing fundamental limitations of existing SSL methods in avian recognition tasks and establishing a new paradigm for efficient self-supervised learning in large-scale ornithological vision applications.
基金supported by the Natural Science Foundation of Inner Mongolia Autonomous Region of China(No.2023QN04011)the National Natural Science Foundation of China(Nos.42307092 and 52279067)+1 种基金Ordos Science and Technology Major Project(No.ZD20232303)Project of Key Laboratory of River and Lake in Inner Mongolia Autonomous Region(No.2022QZBZ0003).
文摘Lacustrine groundwater discharge(LGD)plays an important role in water resources management.Previous studies have focused on LGD process in a single lake,but the differences in LGD process within the same region have not been thoroughly investigated.In this study,multiple tracers(hydrochemistry,𝛿D,𝛿18O and 222Rn)were used to compare mechanisms of LGD in Daihai and Ulansuhai Lake in Inner Mongoli1,Northwest China.The hydrochemical types showed a trend from groundwater to lake water,indicating a hydraulic connection between them.In addition,the𝛿D and𝛿18O values of sediment pore water were between the groundwater and lake water,indicating the LGD processes.The radon mass balance model was used to estimate the average groundwater discharge rates of Daihai and Ulansuhai Lake,which were 2.79 mm/day and 3.02 mm/day,respectively.The total nitrogen(TN),total phosphorus(TP),and fluoride inputs associated with LGD in Daihai Lake accounted for 97.52%,96.59%,and 95.84%of the total inputs,respectively.In contrast,TN,TP and fluoride inputs in Ulansuhai Lake were 53.56%,40.98%,and 36.25%,respectively.This indicates that the pollutant inputs associated with LGD posed a potential threat to the ecological stability of Daihai and Ulansuhai Lake.By comparison,the differences of LGD process and associated pollutant flux were controlled by hydrogeological conditions,lakebed permeability and human activities.This study provides a reference for water resources management in Daihai and Ulansuhai Lake basins while improving the understanding of LGD in the Yellow River basin.
文摘Amidst the growing global emphasis on nuclear safety,the integrity of nuclear reactor systems has garnered attention in the aftermath of consequential events.Moreover,the rapid development of artificial intelligence technology has provided immense opportunities to enhance the safety and economy of nuclear energy.However,data-driven deep learning techniques often lack interpretability,which hinders their applicability in the nuclear energy sector.To address this problem,this study proposes a hybrid data-driven and knowledge-driven artificial intelligence model based on physics-informed neural networks to accurately compute the neutron flux distribution inside a nuclear reactor core.Innovative techniques,such as regional decomposition,intelligent k_(eff)(effective multiplication factor)search,and k_(eff)inversion,have been introduced for the calculation.Furthermore,hyperparameters of the model are automatically optimized using a whale optimization algorithm.A series of computational examples are used to validate the proposed model,demonstrating its applicability,generality,and high accuracy in calculating the neutron flux within the nuclear reactor.The model offers a dependable strategy for computing the neutron flux distribution in nuclear reactors for advanced simulation techniques in the future,including reactor digital twinning.This approach is data-light,requires little to no training data,and still delivers remarkably precise output data.
文摘Background:Medical imaging advancements are constrained by fundamental trade-offs between acquisition speed,radiation dose,and image quality,forcing clinicians to work with noisy,incomplete data.Existing reconstruction methods either compromise on accuracy with iterative algorithms or suffer from limited generalizability with task-specific deep learning approaches.Methods:We present LDM-PIR,a lightweight physics-conditioned diffusion multi-model for medical image reconstruction that addresses key challenges in magnetic resonance imaging(MRI),CT,and low-photon imaging.Unlike traditional iterative methods,which are computationally expensive,or task-specific deep learning approaches lacking generalizability,integrates three innovations.A physics-conditioned diffusion framework that embeds acquisition operators(Fourier/Radon transforms)and noise models directly into the reconstruction process.A multi-model architecture that unifies denoising,inpainting,and super-resolution via shared weight conditioning.A lightweight design(2.1M parameters)enabling rapid inference(0.8s/image on GPU).Through self-supervised fine-tuning with measurement consistency losses adapts to new imaging modalities using fewer annotated samples.Results:Achieves state-of-the-art performance on fastMRI(peak signal-to-noise ratio(PSNR):34.04 for single-coil/31.50 for multi-coil)and Lung Image Database Consortium and Image Database Resource Initiative(28.83 PSNR under Poisson noise).Clinical evaluations demonstrate superior preservation of anatomical structures,with SSIM improvements of 8.8%for single-coil and 4.36%for multi-coil MRI over uDPIR.Conclusion:It offers a flexible,efficient,and scalable solution for medical image reconstruction,addressing the challenges of noise,undersampling,and modality generalization.The model’s lightweight design allows for rapid inference,while its self-supervised fine-tuning capability minimizes reliance on large annotated datasets,making it suitable for real-world clinical applications.
文摘The present study investigates the flow,heat,and mass transfer analysis in the bioconvection of nanofluid containing motile gyrotactic microorganisms through a semi-porous curved oscillatory channel with a magnetic field.These microorganisms produce density gradients by swimming,which induces macroscopic convection flows in the fluid.This procedure improves the mass and heat transfer,illustrating the interaction between biological activity and fluid dynamics.Furthermore,instead of considering traditional Fourier's and Fick's law the energy and concentration equations are developed by incorporating Cattaneo-Christov double diffusion theory.Moreover,to examine the influence of thermophoresis and Brownian diffusions in the fluid we have adopted the Buongiorno nanofluid model.Due to the oscillation of the surface of the channel,the mathematical development of the considered flow problem is obtained in the form of partial differential equations via the curvilinear coordinate system.The convergent series solution of the governing flow equations is obtained after applying the homotopy analysis method(HAM).The effects of different pertinent flow parameters on velocity,motile microorganism density distribution,concentration,pressure,temperature,and skin friction coefficient are examined and discussed in detail with the help of graphs and tables.It is observed during the current study that the density of microorganisms is enhanced for higher values of Reynolds number,Peclet number,radius of curvature variable,and Lewis number.
基金supported by the National Natural Science Foundation of China(Grant No.72161034).
文摘Human motion modeling is a core technology in computer animation,game development,and humancomputer interaction.In particular,generating natural and coherent in-between motion using only the initial and terminal frames remains a fundamental yet unresolved challenge.Existing methods typically rely on dense keyframe inputs or complex prior structures,making it difficult to balance motion quality and plausibility under conditions such as sparse constraints,long-term dependencies,and diverse motion styles.To address this,we propose a motion generation framework based on a frequency-domain diffusion model,which aims to better model complex motion distributions and enhance generation stability under sparse conditions.Our method maps motion sequences to the frequency domain via the Discrete Cosine Transform(DCT),enabling more effective modeling of low-frequency motion structures while suppressing high-frequency noise.A denoising network based on self-attention is introduced to capture long-range temporal dependencies and improve global structural awareness.Additionally,a multi-objective loss function is employed to jointly optimize motion smoothness,pose diversity,and anatomical consistency,enhancing the realism and physical plausibility of the generated sequences.Comparative experiments on the Human3.6M and LaFAN1 datasets demonstrate that our method outperforms state-of-the-art approaches across multiple performance metrics,showing stronger capabilities in generating intermediate motion frames.This research offers a new perspective and methodology for human motion generation and holds promise for applications in character animation,game development,and virtual interaction.
基金Supported by the National Key Research and Development Program of China(No.2023YFC3008200)the Independent Research Project of Southern Marine Science and Engineering Guangdong Laboratory(Zhuhai)(No.SML2022SP505)。
文摘Physics-informed neural networks(PINNs),as a novel artificial intelligence method for solving partial differential equations,are applicable to solve both forward and inverse problems.This study evaluates the performance of PINNs in solving the temperature diffusion equation of the seawater across six scenarios,including forward and inverse problems under three different boundary conditions.Results demonstrate that PINNs achieved consistently higher accuracy with the Dirichlet and Neumann boundary conditions compared to the Robin boundary condition for both forward and inverse problems.Inaccurate weighting of terms in the loss function can reduce model accuracy.Additionally,the sensitivity of model performance to the positioning of sampling points varied between different boundary conditions.In particular,the model under the Dirichlet boundary condition exhibited superior robustness to variations in point positions during the solutions of inverse problems.In contrast,for the Neumann and Robin boundary conditions,accuracy declines when points were sampled from identical positions or at the same time.Subsequently,the Argo observations were used to reconstruct the vertical diffusion of seawater temperature in the north-central Pacific for the applicability of PINNs in the real ocean.The PINNs successfully captured the vertical diffusion characteristics of seawater temperature,reflected the seasonal changes of vertical temperature under different topographic conditions,and revealed the influence of topography on the temperature diffusion coefficient.The PINNs were proved effective in solving the temperature diffusion equation of seawater with limited data,providing a promising technique for simulating or predicting ocean phenomena using sparse observations.
基金supported by the National Natural Science Foundation Joint Fund,No.U22A20309(to PY)the Natural Science Foundation of LiaoningProvince,No.2023-MS-07(to HuL)the Unveiling Key Scientific and Technological Projects of Liaoning Province,No.2021JH1/10400051(to HuL).
文摘Some patients with systemic lupus erythematosus experience neuropsychiatric symptoms.Although magnetic resonance imaging can detect abnormal signals in the white matter of the brain,conventional methods often struggle to accurately capture microstructural changes.Various diffusion models have been used to study white matter in systemic lupus erythematosus;however,comparative analyses of their sensitivity and specificity for detecting microstructural changes remain insufficient.To address this,our team designed a diagnostic trial that used multimodal diffusion imaging techniques to observe white matter microstructural changes in patients with systemic lupus erythematosus who had neuropsychiatric symptoms,with an aim to identify key diagnostic biomarkers for these patients.Patients with active lupus who received treatment at the Department of Rheumatology and Immunology,The First Affiliated Hospital of China Medical University,from September 2023 to March 2024 were recruited.According to the standards of the American College of Rheumatology,patients with systemic lupus erythematosus who had neuropsychiatric symptoms were assigned to the systemic lupus erythematosus group,whereas those without neuropsychiatric symptoms were assigned to the non-systemic lupus erythematosus group.Additionally,healthy volunteers matched by region,sex,and age were recruited as controls.All three groups underwent the same diffusion magnetic resonance imaging examination protocol to compare differences in diffusion parameters.Advanced diffusion imaging models were able to sensitively detect microstructural changes in the white matter fibers of patients with systemic lupus erythematosus who had neuropsychiatric symptoms,with specific diffusion parameters showing significant abnormalities in key brain regions.In the left superior longitudinal fasciculus subregion and the right thalamic radiations of patients with systemic lupus erythematosus who had neuropsychiatric symptoms,we also identified abnormal diffusion characteristics that were clearly correlated with disease activity,suggesting that microstructural changes in these areas may reflect the dynamic process of neuroinflammatory damage.The present study addresses critical challenges in the diagnosis of systemic lupus erythematosus by identifying specific white matter imaging biomarkers and elucidating the association between microstructural damage and clinical manifestations.The main contributions of our study include:1)establishing axial regression probability parameters from mean apparent propagator magnetic resonance imaging as sensitive biomarkers for systemic lupus erythematosus,particularly in the third subregion of the left superior longitudinal fasciculus;2)demonstrating that multimodal diffusion imaging may be superior to conventional diffusion tensor imaging for detecting white matter microstructural abnormalities in patients with systemic lupus erythematosus;and 3)integrating tract-based spatial statistics with clinically relevant analyses to link imaging findings to pathological mechanisms.
文摘The internal flow fields within a three-dimensional inward-tunning combined inlet are extremely complex,especially during the engine mode transition,where the tunnel changes may impact the flow fields significantly.To develop an efficient flow field reconstruction model for this,we present an Improved Conditional Denoising Diffusion Generative Adversarial Network(ICDDGAN),which integrates Conditional Denoising Diffusion Probabilistic Models(CDDPMs)with Style GAN,and introduce a reconstruction discrimination mechanism and dynamic loss weight learning strategy.We establish the Mach number flow field dataset by numerical simulation at various backpressures for the mode transition process from turbine mode to ejector ramjet mode at Mach number 2.5.The proposed ICDDGAN model,given only sparse parameter information,can rapidly generate high-quality Mach number flow fields without a large number of samples for training.The results show that ICDDGAN is superior to CDDGAN in terms of training convergence and stability.Moreover,the interpolation and extrapolation test results during backpressure conditions show that ICDDGAN can accurately and quickly reconstruct Mach number fields at various tunnel slice shapes,with a Structural Similarity Index Measure(SSIM)of over 0.96 and a Mean-Square Error(MSE)of 0.035%to actual flow fields,reducing time costs by 7-8 orders of magnitude compared to Computational Fluid Dynamics(CFD)calculations.This can provide an efficient means for rapid computation of complex flow fields.
