Dear Editor,The letter proposes a tensor low-rank orthogonal compression(TLOC)model for a convolutional neural network(CNN),which facilitates its efficient and highly-accurate low-rank representation.Model compression...Dear Editor,The letter proposes a tensor low-rank orthogonal compression(TLOC)model for a convolutional neural network(CNN),which facilitates its efficient and highly-accurate low-rank representation.Model compression is crucial for deploying deep neural network(DNN)models on resource-constrained embedded devices.展开更多
Dear Editor,This letter addresses the critical challenge of preserving privacy in graph learning without compromising on data utility.Differential privacy(DP)is emerging as an effective method for privacy-preserving g...Dear Editor,This letter addresses the critical challenge of preserving privacy in graph learning without compromising on data utility.Differential privacy(DP)is emerging as an effective method for privacy-preserving graph learning.However,its application often diminishes data utility,especially for nodes with fewer neighbors in graph neural networks(GNNs).展开更多
Near-infrared image sensors are widely used in fields such as material identification,machine vision,and autonomous driving.Lead sulfide colloidal quantum dot-based infrared photodiodes can be integrated with sil⁃icon...Near-infrared image sensors are widely used in fields such as material identification,machine vision,and autonomous driving.Lead sulfide colloidal quantum dot-based infrared photodiodes can be integrated with sil⁃icon-based readout circuits in a single step.Based on this,we propose a photodiode based on an n-i-p structure,which removes the buffer layer and further simplifies the manufacturing process of quantum dot image sensors,thus reducing manufacturing costs.Additionally,for the noise complexity in quantum dot image sensors when capturing images,traditional denoising and non-uniformity methods often do not achieve optimal denoising re⁃sults.For the noise and stripe-type non-uniformity commonly encountered in infrared quantum dot detector imag⁃es,a network architecture has been developed that incorporates multiple key modules.This network combines channel attention and spatial attention mechanisms,dynamically adjusting the importance of feature maps to en⁃hance the ability to distinguish between noise and details.Meanwhile,the residual dense feature fusion module further improves the network's ability to process complex image structures through hierarchical feature extraction and fusion.Furthermore,the pyramid pooling module effectively captures information at different scales,improv⁃ing the network's multi-scale feature representation ability.Through the collaborative effect of these modules,the network can better handle various mixed noise and image non-uniformity issues.Experimental results show that it outperforms the traditional U-Net network in denoising and image correction tasks.展开更多
Speckle-based optical cryptosystems are promising technologies for information security.However,existing techniques mostly rely on digital decryption,resulting in computational delay and undermining the high-speed adv...Speckle-based optical cryptosystems are promising technologies for information security.However,existing techniques mostly rely on digital decryption,resulting in computational delay and undermining the high-speed advantage of optical encryption.Moreover,conventional neural networks are typically effective only on images from the same distribution as the training datasets,limiting their general applicability.In this paper,we propose an all-optical high-speed decryption scheme for real-time recovery of speckle-encoded ciphertexts.By constructing a physics-informed diffractive neural network that approximates the inverse transmission matrix of the scattering medium,secret images can be directly reconstructed from speckle fields without optoelectronic conversion or post-processing.The network is trained with only 2048 samples from the MNIST dataset.Its transfer learning capability is validated across three out-of-distribution datasets,with decrypted images achieving a Pearson correlation coefficient of 0.82 and a structural similarity index measure of 0.75,demonstrating excellent transfer learning capability.For the first time,to our knowledge,this scheme simultaneously overcomes the bottlenecks of decryption delay and limited network generalizability in conventional speckle-based cryptosystems,achieving real-time image decryption with strong transferability.It provides a new pathway for developing low-power,real-time,and broadly applicable optical encryption systems,demonstrating significant potential for applications in high-speed security optical communications.展开更多
Rapid quantification of seismic-induced damage immediately following an earthquake is critical for determining whether a structure is safe for continued occupation or requires evacuation.This study proposes a novel da...Rapid quantification of seismic-induced damage immediately following an earthquake is critical for determining whether a structure is safe for continued occupation or requires evacuation.This study proposes a novel damage identification method that utilizes limited strain data points,significantly reducing installation,maintenance,and data analysis costs compared to traditional distributed sensor networks.The approach integrates finite element(FE)modeling to generate capacity curves through pushover analysis,incorporates noise-augmented datasets for Artificial Neural Network(ANN)training,and classifies structural conditions into four damage levels:Operational(OP),Immediate Occupancy(IO),Life Safety(LS),and Collapse Prevention(CP).To evaluate the method’s accuracy and efficiency,it was applied to two reinforced concrete(RC)frames;a single-story frame tested experimentally under cyclic loading and a three-story frame analyzed under various lateral load patterns.Strain data from selected beam and column ends were used as ANN inputs,while the corresponding damage classes served as outputs.Confusion matrix results demonstrated high true positive rates(>85%for the single-story and>90%for the three-story frame),even with a reduced number of sensors.The model also exhibited strong robustness to White Gaussian Noise(SNR=2.5-5 dB)and generalized effectively to nonlinear time-history analyses under scaled ground motions(PGA=0.1-1.0 g).Feature selection using the MRMR and ANOVA algorithms further enhanced computational efficiency.Overall,the proposed ANN-based framework has strong potential for real-time structural health monitoring applications.展开更多
This paper introduces a fuzzy C-means-based pooling layer for convolutional neural networks that explicitly models local uncertainty and ambiguity.Conventional pooling operations,such as max and average,apply rigid ag...This paper introduces a fuzzy C-means-based pooling layer for convolutional neural networks that explicitly models local uncertainty and ambiguity.Conventional pooling operations,such as max and average,apply rigid aggregation and often discard fine-grained boundary information.In contrast,our method computes soft membershipswithin each receptive field and aggregates cluster-wise responses throughmembership-weighted pooling,thereby preserving informative structure while reducing dimensionality.Being differentiable,the proposed layer operates as standard two-dimensional pooling.We evaluate our approach across various CNN backbones and open datasets,including CIFAR-10/100,STL-10,LFW,and ImageNette,and further probe small training set restrictions on MNIST and Fashion-MNIST.In these settings,the proposed pooling consistently improves accuracy and weighted F1 over conventional baselines,with particularly strong gains when training data are scarce.Even with less than 1%of the training set,ourmethodmaintains reliable performance,indicating improved sample efficiency and robustness to noisy or ambiguous local patterns.Overall,integrating soft memberships into the pooling operator provides a practical and generalizable inductive bias that enhances robustness and generalization in modern CNN pipelines.展开更多
Distributed Denial of Service(DDoS)attacks are one of the severe threats to network infrastructure,sometimes bypassing traditional diagnosis algorithms because of their evolving complexity.PresentMachine Learning(ML)t...Distributed Denial of Service(DDoS)attacks are one of the severe threats to network infrastructure,sometimes bypassing traditional diagnosis algorithms because of their evolving complexity.PresentMachine Learning(ML)techniques for DDoS attack diagnosis normally apply network traffic statistical features such as packet sizes and inter-arrival times.However,such techniques sometimes fail to capture complicated relations among various traffic flows.In this paper,we present a new multi-scale ensemble strategy given the Graph Neural Networks(GNNs)for improving DDoS detection.Our technique divides traffic into macro-and micro-level elements,letting various GNN models to get the two corase-scale anomalies and subtle,stealthy attack models.Through modeling network traffic as graph-structured data,GNNs efficiently learn intricate relations among network entities.The proposed ensemble learning algorithm combines the results of several GNNs to improve generalization,robustness,and scalability.Extensive experiments on three benchmark datasets—UNSW-NB15,CICIDS2017,and CICDDoS2019—show that our approach outperforms traditional machine learning and deep learning models in detecting both high-rate and low-rate(stealthy)DDoS attacks,with significant improvements in accuracy and recall.These findings demonstrate the suggested method’s applicability and robustness for real-world implementation in contexts where several DDoS patterns coexist.展开更多
Physics-informed neural networks(PINNs)have emerged as a promising class of scientific machine learning techniques that integrate governing physical laws into neural network training.Their ability to enforce different...Physics-informed neural networks(PINNs)have emerged as a promising class of scientific machine learning techniques that integrate governing physical laws into neural network training.Their ability to enforce differential equations,constitutive relations,and boundary conditions within the loss function provides a physically grounded alternative to traditional data-driven models,particularly for solid and structural mechanics,where data are often limited or noisy.This review offers a comprehensive assessment of recent developments in PINNs,combining bibliometric analysis,theoretical foundations,application-oriented insights,and methodological innovations.A biblio-metric survey indicates a rapid increase in publications on PINNs since 2018,with prominent research clusters focused on numerical methods,structural analysis,and forecasting.Building upon this trend,the review consolidates advance-ments across five principal application domains,including forward structural analysis,inverse modeling and parameter identification,structural and topology optimization,assessment of structural integrity,and manufacturing processes.These applications are propelled by substantial methodological advancements,encompassing rigorous enforcement of boundary conditions,modified loss functions,adaptive training,domain decomposition strategies,multi-fidelity and transfer learning approaches,as well as hybrid finite element–PINN integration.These advances address recurring challenges in solid mechanics,such as high-order governing equations,material heterogeneity,complex geometries,localized phenomena,and limited experimental data.Despite remaining challenges in computational cost,scalability,and experimental validation,PINNs are increasingly evolving into specialized,physics-aware tools for practical solid and structural mechanics applications.展开更多
Advances in Alzheimer's disease(AD)research have deepened our understanding,yet the mechanisms driving its progression remain unclear.Although a range of in vivo biomarkers is now available(e.g.,measurements of am...Advances in Alzheimer's disease(AD)research have deepened our understanding,yet the mechanisms driving its progression remain unclear.Although a range of in vivo biomarkers is now available(e.g.,measurements of amyloidbeta(Aβ)and ta u accumulation-the molecular hallmarks of AD-structural magnetic resonance imaging(MRI),assessments of brain metabolism,and,more recently,blood-based markers),a definitive diagnosis of AD continues to be challenging.For example,Frisoni et al.展开更多
critical for guiding treatment and improving patient outcomes.Traditional molecular subtyping via immuno-histochemistry(IHC)test is invasive,time-consuming,and may not fully represent tumor heterogeneity.This study pr...critical for guiding treatment and improving patient outcomes.Traditional molecular subtyping via immuno-histochemistry(IHC)test is invasive,time-consuming,and may not fully represent tumor heterogeneity.This study proposes a non-invasive approach using digital mammography images and deep learning algorithm for classifying breast cancer molecular subtypes.Four pretrained models,including two Convolutional Neural Networks(MobileNet_V3_Large and VGG-16)and two Vision Transformers(ViT_B_16 and ViT_Base_Patch16_Clip_224)were fine-tuned to classify images into HER2-enriched,Luminal,Normal-like,and Triple Negative subtypes.Hyperparameter tuning,including learning rate adjustment and layer freezing strategies,was applied to optimize performance.Among the evaluated models,ViT_Base_Patch16_Clip_224 achieved the highest test accuracy(94.44%),with equally high precision,recall,and F1-score of 0.94,demonstrating excellent generalization.MobileNet_V3_Large achieved the same accuracy but showed less training stability.In contrast,VGG-16 recorded the lowest performance,indicating a limitation in its generalizability for this classification task.The study also highlighted the superior performance of the Vision Transformer models over CNNs,particularly due to their ability to capture global contextual features and the benefit of CLIP-based pretraining in ViT_Base_Patch16_Clip_224.To enhance clinical applicability,a graphical user interface(GUI)named“BCMS Dx”was developed for streamlined subtype prediction.Deep learning applied to mammography has proven effective for accurate and non-invasive molecular subtyping.The proposed Vision Transformer-based model and supporting GUI offer a promising direction for augmenting diagnostic workflows,minimizing the need for invasive procedures,and advancing personalized breast cancer management.展开更多
The thermal conductivity of nanofluids is an important property that influences the heat transfer capabilities of nanofluids.Researchers rely on experimental investigations to explore nanofluid properties,as it is a n...The thermal conductivity of nanofluids is an important property that influences the heat transfer capabilities of nanofluids.Researchers rely on experimental investigations to explore nanofluid properties,as it is a necessary step before their practical application.As these investigations are time and resource-consuming undertakings,an effective prediction model can significantly improve the efficiency of research operations.In this work,an Artificial Neural Network(ANN)model is developed to predict the thermal conductivity of metal oxide water-based nanofluid.For this,a comprehensive set of 691 data points was collected from the literature.This dataset is split into training(70%),validation(15%),and testing(15%)and used to train the ANN model.The developed model is a backpropagation artificial neural network with a 4–12–1 architecture.The performance of the developed model shows high accuracy with R values above 0.90 and rapid convergence.It shows that the developed ANN model accurately predicts the thermal conductivity of nanofluids.展开更多
The flight envelope of Air Turbo Rocket(ATR)engines is broader compared to conventional aero-engines,and designing a full-envelope controller using traditional methods poses significant challenges due to a burdensome ...The flight envelope of Air Turbo Rocket(ATR)engines is broader compared to conventional aero-engines,and designing a full-envelope controller using traditional methods poses significant challenges due to a burdensome design process.To address this issue,this paper proposes a self-learning neural network controller design method based on Reinforcement Learning(RL).Additionally,a method for predictive compensation and stability rewards is proposed to reduce the system oscillation caused by actuator delay.This approach simplifies the actuator to a firstorder inertial element exhibiting pure delay.A simulation environment for the ATR engineactuator system is first established.Based on this environment,a self-learning neural network controller using a predictive compensator and the Proximal Policy Optimization(PPO)algorithm is then developed.Furthermore,the temporal difference signals from the controller output are integrated into the reward function to enhance system stability.The proposed method is validated through numerical simulations and semi-physical experiments.The numerical simulation results demonstrate that the proposed method increases the system's tolerance to delays from 20 ms to 400 ms.Under an actuator delay of 400 ms,the average steady-state error remains less than0.1%,the overshoot is limited to 1%,and the settling time does not exceed 3 s.Moreover,compared to the traditional method,the proposed method exhibits higher adaptability to model errors and variations in flight conditions.In the conducted semi-physical simulation experiments,the proposed method achieves stable control of a real electric pump.展开更多
Accurate state of health(SOH)estimation is essential for the safe and reliable operation of lithium-ion batteries.However,existing methods face significant challenges,primarily because they rely on complete charge–di...Accurate state of health(SOH)estimation is essential for the safe and reliable operation of lithium-ion batteries.However,existing methods face significant challenges,primarily because they rely on complete charge–discharge cycles and fixed-form physical constraints,which limit adaptability to different chemistries and real-world conditions.To address these issues,this study proposes an approach that extracts features from segmented state of charge(SOC)intervals and integrates them into an enhanced physics-informed neural network(PINN).Specifically,voltage data within the 25%–75%SOC range during charging are used to derive statistical,time–frequency,and mechanism-based features that capture degradation trends.A hybrid PINN-Lasso-Transformer-BiLSTM architecture is developed,where Lasso regression enables sparse feature selection,and a nonlinear empirical degradation model is embedded as a learnable physical term within a dynamically scaled composite loss.This design adaptively balances data-driven accuracy with physical consistency,thereby enhancing estimation precision,robustness,and generalization.The results show that the proposed method outperforms conventional neural networks across four battery chemistries,achieving root mean square error and mean absolute error below 1%.Notably,features from partial charging segments exhibit higher robustness than those from full cycles.Furthermore,the model maintains strong performance under high temperatures and demonstrates excellent generalization capacity in transfer learning across chemistries,temperatures,and C-rates.This work establishes a scalable and interpretable solution for accurate SOH estimation under diverse practical operating conditions.展开更多
The adjoint method is widely used in gradient-based optimization with high-dimensional design variables.However,the cost of solving the adjoint equations in each iteration is comparable to that of solving the flow fie...The adjoint method is widely used in gradient-based optimization with high-dimensional design variables.However,the cost of solving the adjoint equations in each iteration is comparable to that of solving the flow field,resulting in expensive computational costs.To improve the efficiency of solving adjoint equations,we propose a physics-constrained graph neural networks for solving adjoint equations,named ADJ-PCGN.ADJ-PCGN establishes a mapping relationship between flow characteristics and adjoint vector based on data,serving as a replacement for the computationally expensive numerical solution of adjoint equations.A physics-based graph structure and message-passing mechanism are designed to endow its strong fitting and generalization capabilities.Taking transonic drag reduction and maximum lift-drag ratio of the airfoil as examples,results indicate that ADJ-PCGN attains a similar optimal shape as the classical direct adjoint loop method.In addition,ADJ-PCGN demonstrates strong generalization capabilities across different mesh topologies,mesh densities,and out-of-distribution conditions.It holds the potential to become a universal model for aerodynamic shape optimization involving states,geometries,and meshes.展开更多
It is well known that aluminum and copper exhibit structural phase transformations in quasi-static and dynamic measurements,including shock wave loading.However,the dependence of phase transformations in a wide range ...It is well known that aluminum and copper exhibit structural phase transformations in quasi-static and dynamic measurements,including shock wave loading.However,the dependence of phase transformations in a wide range of crystallographic directions of shock loading has not been revealed.In this work,we calculated the shock Hugoniot for aluminum and copper in different crystallographic directions([100],[110],[111],[112],[102],[114],[123],[134],[221]and[401])of shock compression using molecular dynamics(MD)simulations.The results showed a high pressure(>160 GPa for Cu and>40 GPa for Al)of the FCC-to-BCC transition.In copper,different characteristics of the phase transition are observed depending on the loading direction with the[100]compression direction being the weakest.The FCC-to-BCC transition for copper is in the range of 150–220 GPa,which is consistent with the existing experimental data.Due to the high transition pressure,the BCC phase transition in copper competes with melting.In aluminum,the FCC-to-BCC transition is observed for all studied directions at pressures between 40 and 50 GPa far beyond the melting.In all considered cases we observe the coexistence of HCP and BCC phases during the FCC-to-BCC transition,which is consistent with the experimental data and atomistic calculations;this HCP phase forms in the course of accompanying plastic deformation with dislocation activity in the parent FCC phase.The plasticity incipience is also anisotropic in bothmetals,which is due to the difference in the projections of stress on the slip plane for different orientations of the FCC crystal.MD modeling results demonstrate a strong dependence of the FCC-to-BCC transition on the crystallographic direction,in which the material is loaded in the copper crystals.However,MD simulations data can only be obtained for specific points in the stereographic direction space;therefore,for more comprehensive understanding of the phase transition process,a feed-forward neural network was trained using MD modeling data.The trained machine learning model allowed us to construct continuous stereographic maps of phase transitions as a function of stress in the shock-compressed state of metal.Due to appearance and growth of multiple centers of new phase,the FCC-to-BCC transition leads to formation of a polycrystalline structure from the parent single crystal.展开更多
Recently,the zeroing neural network(ZNN)has demonstrated remarkable effectiveness in tackling time-varying problems,delivering robust performance across both noise-free and noisy environments.However,existing ZNN mode...Recently,the zeroing neural network(ZNN)has demonstrated remarkable effectiveness in tackling time-varying problems,delivering robust performance across both noise-free and noisy environments.However,existing ZNN models are limited in their ability to actively suppress noise,which constrains their robustness and precision in solving time-varying problems.This paper introduces a novel active noise rejection ZNN(ANR-ZNN)design that enhances noise suppression by integrating computational error dynamics and harmonic behaviour.Through rigorous theoretical analysis,we demonstrate that the proposed ANR-ZNN maintains robust convergence in computational error performance under environmental noise.As a case study,the ANR-ZNN model is specifically applied to time-varying matrix inversion.Comprehensive computer simulations and robotic experiments further validate the ANR-ZNN's effectiveness,emphasising the proposed design's superiority and potential for solving time-varying problems.展开更多
There is a need for accurate prediction of heat and mass transfer in aerodynamically designed,non-Newtonian nanofluids across aerodynamically designed,high-flux biomedical micro-devices for thermal management and reac...There is a need for accurate prediction of heat and mass transfer in aerodynamically designed,non-Newtonian nanofluids across aerodynamically designed,high-flux biomedical micro-devices for thermal management and reactive coating processes,but existing work is not uncharacteristically remiss regarding viscoelasticity,radiative heating,viscous dissipation,and homogeneous–heterogeneous reactions within a single scheme that is calibrated.This research investigates the flow of Williamson nanofluid across a dynamically wedged surface under conditions that include viscous dissipation,thermal radiation,and homogeneous-heterogeneous reactions.The paper develops a detailed mathematical approach that utilizes boundary layers to transform partial differential equations into ordinary differential equations using similarity transformations.RK4 is the technique for gaining numerical solutions,but with the addition of ANNs,there is an improvement in prediction accuracy and computational efficiency.The study investigates the influence of wedge angle parameter,along with Weissenberg number,thermal radiation parameter and Brownian motion parameter,and Schmidt number,on velocity distribution,temperature distribution,and concentra-tion distribution.Enhanced Weissenberg numbers enhance viscoelastic responses that modify velocity patterns,but radiation parameters and thermophoresis have key impacts on thermal transfer phenomena.This research develops findings that are of enormous application in aerospace,biomedical(artificial hearts and drug delivery),and industrial cooling technology applications.New findings on non-Newtonian nanofluids under full flow systems are included in this work to enhance heat transfer methods in novel fluid-based systems.展开更多
Since the advent of smart contracts,security vulnerabilities have remained a persistent challenge,compromsing both the reliability of contract execution and the overall stability of the virtual currency market.Consequ...Since the advent of smart contracts,security vulnerabilities have remained a persistent challenge,compromsing both the reliability of contract execution and the overall stability of the virtual currency market.Consequently,the academic community has devoted increasing attention to these security risks.However,conventional approaches to vulnerability detection frequently exhibit limited accuracy.To address this limitation,the present study introduces a novel vulnerability detection framework called GNNSE that integrates symbolic execution with graph neural networks(GNNs).The proposedmethod first constructs semantic graphs to comprehensively capture the control flow and data flow dependencies within smart contracts.These graphs are subsequently processed using GNNs to efficiently identify contracts with a high likelihood of vulnerabilities.For these high-risk contracts,symbolic execution is employed to perform fine-grained,path-level analysis,thereby improving overall detection precision.Experimental results on a dataset comprising 10,079 contracts demonstrate that the proposed method achieves detection precisions of 93.58% for reentrancy vulnerabilities and 92.73% for timestamp-dependent vulnerabilities.展开更多
In contemporary computer vision,convolutional neural networks(CNNs)and vision transformers(ViTs)represent the two primary architectural paradigms for image recognition.While both approaches have been widely adopted in...In contemporary computer vision,convolutional neural networks(CNNs)and vision transformers(ViTs)represent the two primary architectural paradigms for image recognition.While both approaches have been widely adopted in medical imaging applications,they operate based on fundamentally different computational principles.This report attempts to provide brief application notes on ViTs and CNNs,particularly focusing on scenarios that guide the selection of one architecture over the other in practical medical implementations.Generally,CNNs rely on convolutional kernels,localized receptive fields,and weight sharing,enabling efficient hierarchical feature extraction.These properties contribute to strong performance in detecting spatially constrained patterns such as textures,edges,and anatomical boundaries,while maintaining relatively low computational requirements.ViTs,on the other hand,decompose images into smaller segments referred to as tokens and employ self-attention mechanisms to model relationships across the entire image.This global modeling capability allows ViTs to capture long-range dependencies that may be difficult for convolution-based architectures to learn.However,ViTs typically achieve optimal performance when trained on extremely large datasets or when supported by extensive pretraining,as their reduced inductive bias requires greater data exposure to learn robust representations.This report briefly examines the architectural structure,underlying mathematical foundations,and relative performance characteristics of CNNs and ViTs,drawing upon recent findings from contemporary research.Emphasis is placed on understanding how differences in data availability,computational resources,and task requirements influence model effectiveness across medical imaging domains.Most importantly,the report serves as a concise application guide for practitioners seeking informed implementation decisions between these two influential deep learning frameworks.展开更多
The development of β-titanium alloys with bone-mimicking elastic moduli remains a significant challenge.Although machine learning has the potential to accelerate alloy discovery,traditional methods often face data li...The development of β-titanium alloys with bone-mimicking elastic moduli remains a significant challenge.Although machine learning has the potential to accelerate alloy discovery,traditional methods often face data limitations such as sparsity,compositional discontinuity,and feature heterogeneity,leading to overfitting and restricting the exploration of novel compositional spaces.In this study,we introduce a domain-adversarial neural network framework that balances predictive accuracy with the generalization ability of unexplored composition space through integrated feature alignment and adversarial training.Using this approach,we successfully developed a non-intuitiveβ-Ti alloy with an ultra-low elastic modulus of 28±3 GPa,providing new insights beyond conventionally designed biomedical titanium alloys.This work establishes a screening framework for materials discovery in small-sample data spaces,with broad implications for the design of biomedical and other alloy systems.展开更多
基金supported by the Science and Technology Innovation Key R&D Program of Chongqing(CSTB2025TIAD-STX0032)National Key Research and Development Program of China(2024YFF0908200)+1 种基金the Chongqing Technology Innovation and Application Development Special Key Project(CSTB2024TIAD-KPX0018)the Southwest University Graduate Student Research Innovation(SWUB24051)。
文摘Dear Editor,The letter proposes a tensor low-rank orthogonal compression(TLOC)model for a convolutional neural network(CNN),which facilitates its efficient and highly-accurate low-rank representation.Model compression is crucial for deploying deep neural network(DNN)models on resource-constrained embedded devices.
基金supported by the National Key Research and Development Program of China(2023YFF0612900,2023YFF0612902)the Natural Science Foundation of Beijing,China(4254086)+3 种基金the National Natural Science Foundation of China(62472032)the Open Project Funding of Key Laboratory of Mobile Application Innovation and Governance Technology,Ministry of Industry and Information Technology(2023IFS080601-K)the Beijing Institute of Technology Research Fund Program for Young Scholarsthe Young Elite Scientists Sponsorship Program by CAST(2023QNRC001)。
文摘Dear Editor,This letter addresses the critical challenge of preserving privacy in graph learning without compromising on data utility.Differential privacy(DP)is emerging as an effective method for privacy-preserving graph learning.However,its application often diminishes data utility,especially for nodes with fewer neighbors in graph neural networks(GNNs).
基金Supported by the National key research and development program in the 14th five year plan 2021YFA1200700)the National Natural Science Foundation of China(62535018,62431025,62561160113)the Natural Science Foundation of Shanghai(23ZR1473400).
文摘Near-infrared image sensors are widely used in fields such as material identification,machine vision,and autonomous driving.Lead sulfide colloidal quantum dot-based infrared photodiodes can be integrated with sil⁃icon-based readout circuits in a single step.Based on this,we propose a photodiode based on an n-i-p structure,which removes the buffer layer and further simplifies the manufacturing process of quantum dot image sensors,thus reducing manufacturing costs.Additionally,for the noise complexity in quantum dot image sensors when capturing images,traditional denoising and non-uniformity methods often do not achieve optimal denoising re⁃sults.For the noise and stripe-type non-uniformity commonly encountered in infrared quantum dot detector imag⁃es,a network architecture has been developed that incorporates multiple key modules.This network combines channel attention and spatial attention mechanisms,dynamically adjusting the importance of feature maps to en⁃hance the ability to distinguish between noise and details.Meanwhile,the residual dense feature fusion module further improves the network's ability to process complex image structures through hierarchical feature extraction and fusion.Furthermore,the pyramid pooling module effectively captures information at different scales,improv⁃ing the network's multi-scale feature representation ability.Through the collaborative effect of these modules,the network can better handle various mixed noise and image non-uniformity issues.Experimental results show that it outperforms the traditional U-Net network in denoising and image correction tasks.
基金supported by the Guangdong Major Project of Basic Research(Grant No.2020B0301030009)the National Natural Science Foundation of China(Grant Nos.12174204,12174203,12074203,62335012,and 62435010)+5 种基金the Natural Science Foundation of Guangdong Province(Grant No.2023A1515012888)the Science and Technology Innovation Commission of Shenzhen(Grant Nos.JCYJ20220818101417039 and JCYJ20241202124428038)the Medical-Engineering Interdisciplinary Research Foundation of Shenzhen University(Grant No.86901/00000311)the Scientific Instrument Developing Project of Shenzhen University(Grant No.2023YQ001)the Shenzhen University 2035 Initiative(Grant No.2023B004)the Key R&D Program of Zhejiang(Grant No.30003AA240100)。
文摘Speckle-based optical cryptosystems are promising technologies for information security.However,existing techniques mostly rely on digital decryption,resulting in computational delay and undermining the high-speed advantage of optical encryption.Moreover,conventional neural networks are typically effective only on images from the same distribution as the training datasets,limiting their general applicability.In this paper,we propose an all-optical high-speed decryption scheme for real-time recovery of speckle-encoded ciphertexts.By constructing a physics-informed diffractive neural network that approximates the inverse transmission matrix of the scattering medium,secret images can be directly reconstructed from speckle fields without optoelectronic conversion or post-processing.The network is trained with only 2048 samples from the MNIST dataset.Its transfer learning capability is validated across three out-of-distribution datasets,with decrypted images achieving a Pearson correlation coefficient of 0.82 and a structural similarity index measure of 0.75,demonstrating excellent transfer learning capability.For the first time,to our knowledge,this scheme simultaneously overcomes the bottlenecks of decryption delay and limited network generalizability in conventional speckle-based cryptosystems,achieving real-time image decryption with strong transferability.It provides a new pathway for developing low-power,real-time,and broadly applicable optical encryption systems,demonstrating significant potential for applications in high-speed security optical communications.
基金funded by UTM Fundamental Research Grant(PY/2024/01221,Cost centre no.:Q.J130000.3822.23H73)HiCoE Grant Scheme(Cost centre no.:R.J130000.7822.4J738)。
文摘Rapid quantification of seismic-induced damage immediately following an earthquake is critical for determining whether a structure is safe for continued occupation or requires evacuation.This study proposes a novel damage identification method that utilizes limited strain data points,significantly reducing installation,maintenance,and data analysis costs compared to traditional distributed sensor networks.The approach integrates finite element(FE)modeling to generate capacity curves through pushover analysis,incorporates noise-augmented datasets for Artificial Neural Network(ANN)training,and classifies structural conditions into four damage levels:Operational(OP),Immediate Occupancy(IO),Life Safety(LS),and Collapse Prevention(CP).To evaluate the method’s accuracy and efficiency,it was applied to two reinforced concrete(RC)frames;a single-story frame tested experimentally under cyclic loading and a three-story frame analyzed under various lateral load patterns.Strain data from selected beam and column ends were used as ANN inputs,while the corresponding damage classes served as outputs.Confusion matrix results demonstrated high true positive rates(>85%for the single-story and>90%for the three-story frame),even with a reduced number of sensors.The model also exhibited strong robustness to White Gaussian Noise(SNR=2.5-5 dB)and generalized effectively to nonlinear time-history analyses under scaled ground motions(PGA=0.1-1.0 g).Feature selection using the MRMR and ANOVA algorithms further enhanced computational efficiency.Overall,the proposed ANN-based framework has strong potential for real-time structural health monitoring applications.
文摘This paper introduces a fuzzy C-means-based pooling layer for convolutional neural networks that explicitly models local uncertainty and ambiguity.Conventional pooling operations,such as max and average,apply rigid aggregation and often discard fine-grained boundary information.In contrast,our method computes soft membershipswithin each receptive field and aggregates cluster-wise responses throughmembership-weighted pooling,thereby preserving informative structure while reducing dimensionality.Being differentiable,the proposed layer operates as standard two-dimensional pooling.We evaluate our approach across various CNN backbones and open datasets,including CIFAR-10/100,STL-10,LFW,and ImageNette,and further probe small training set restrictions on MNIST and Fashion-MNIST.In these settings,the proposed pooling consistently improves accuracy and weighted F1 over conventional baselines,with particularly strong gains when training data are scarce.Even with less than 1%of the training set,ourmethodmaintains reliable performance,indicating improved sample efficiency and robustness to noisy or ambiguous local patterns.Overall,integrating soft memberships into the pooling operator provides a practical and generalizable inductive bias that enhances robustness and generalization in modern CNN pipelines.
文摘Distributed Denial of Service(DDoS)attacks are one of the severe threats to network infrastructure,sometimes bypassing traditional diagnosis algorithms because of their evolving complexity.PresentMachine Learning(ML)techniques for DDoS attack diagnosis normally apply network traffic statistical features such as packet sizes and inter-arrival times.However,such techniques sometimes fail to capture complicated relations among various traffic flows.In this paper,we present a new multi-scale ensemble strategy given the Graph Neural Networks(GNNs)for improving DDoS detection.Our technique divides traffic into macro-and micro-level elements,letting various GNN models to get the two corase-scale anomalies and subtle,stealthy attack models.Through modeling network traffic as graph-structured data,GNNs efficiently learn intricate relations among network entities.The proposed ensemble learning algorithm combines the results of several GNNs to improve generalization,robustness,and scalability.Extensive experiments on three benchmark datasets—UNSW-NB15,CICIDS2017,and CICDDoS2019—show that our approach outperforms traditional machine learning and deep learning models in detecting both high-rate and low-rate(stealthy)DDoS attacks,with significant improvements in accuracy and recall.These findings demonstrate the suggested method’s applicability and robustness for real-world implementation in contexts where several DDoS patterns coexist.
基金funded by National Research Council of Thailand(contract No.N42A671047).
文摘Physics-informed neural networks(PINNs)have emerged as a promising class of scientific machine learning techniques that integrate governing physical laws into neural network training.Their ability to enforce differential equations,constitutive relations,and boundary conditions within the loss function provides a physically grounded alternative to traditional data-driven models,particularly for solid and structural mechanics,where data are often limited or noisy.This review offers a comprehensive assessment of recent developments in PINNs,combining bibliometric analysis,theoretical foundations,application-oriented insights,and methodological innovations.A biblio-metric survey indicates a rapid increase in publications on PINNs since 2018,with prominent research clusters focused on numerical methods,structural analysis,and forecasting.Building upon this trend,the review consolidates advance-ments across five principal application domains,including forward structural analysis,inverse modeling and parameter identification,structural and topology optimization,assessment of structural integrity,and manufacturing processes.These applications are propelled by substantial methodological advancements,encompassing rigorous enforcement of boundary conditions,modified loss functions,adaptive training,domain decomposition strategies,multi-fidelity and transfer learning approaches,as well as hybrid finite element–PINN integration.These advances address recurring challenges in solid mechanics,such as high-order governing equations,material heterogeneity,complex geometries,localized phenomena,and limited experimental data.Despite remaining challenges in computational cost,scalability,and experimental validation,PINNs are increasingly evolving into specialized,physics-aware tools for practical solid and structural mechanics applications.
文摘Advances in Alzheimer's disease(AD)research have deepened our understanding,yet the mechanisms driving its progression remain unclear.Although a range of in vivo biomarkers is now available(e.g.,measurements of amyloidbeta(Aβ)and ta u accumulation-the molecular hallmarks of AD-structural magnetic resonance imaging(MRI),assessments of brain metabolism,and,more recently,blood-based markers),a definitive diagnosis of AD continues to be challenging.For example,Frisoni et al.
基金funded by the Ministry of Higher Education(MoHE)Malaysia through the Fundamental Research Grant Scheme—Early Career Researcher(FRGS-EC),grant number FRGSEC/1/2024/ICT02/UNIMAP/02/8.
文摘critical for guiding treatment and improving patient outcomes.Traditional molecular subtyping via immuno-histochemistry(IHC)test is invasive,time-consuming,and may not fully represent tumor heterogeneity.This study proposes a non-invasive approach using digital mammography images and deep learning algorithm for classifying breast cancer molecular subtypes.Four pretrained models,including two Convolutional Neural Networks(MobileNet_V3_Large and VGG-16)and two Vision Transformers(ViT_B_16 and ViT_Base_Patch16_Clip_224)were fine-tuned to classify images into HER2-enriched,Luminal,Normal-like,and Triple Negative subtypes.Hyperparameter tuning,including learning rate adjustment and layer freezing strategies,was applied to optimize performance.Among the evaluated models,ViT_Base_Patch16_Clip_224 achieved the highest test accuracy(94.44%),with equally high precision,recall,and F1-score of 0.94,demonstrating excellent generalization.MobileNet_V3_Large achieved the same accuracy but showed less training stability.In contrast,VGG-16 recorded the lowest performance,indicating a limitation in its generalizability for this classification task.The study also highlighted the superior performance of the Vision Transformer models over CNNs,particularly due to their ability to capture global contextual features and the benefit of CLIP-based pretraining in ViT_Base_Patch16_Clip_224.To enhance clinical applicability,a graphical user interface(GUI)named“BCMS Dx”was developed for streamlined subtype prediction.Deep learning applied to mammography has proven effective for accurate and non-invasive molecular subtyping.The proposed Vision Transformer-based model and supporting GUI offer a promising direction for augmenting diagnostic workflows,minimizing the need for invasive procedures,and advancing personalized breast cancer management.
基金supported by Basic Science Research Program through the National Research Foundation of Korea(NRF)funded by the Ministry of Education(2021R1A6A1A10044950).
文摘The thermal conductivity of nanofluids is an important property that influences the heat transfer capabilities of nanofluids.Researchers rely on experimental investigations to explore nanofluid properties,as it is a necessary step before their practical application.As these investigations are time and resource-consuming undertakings,an effective prediction model can significantly improve the efficiency of research operations.In this work,an Artificial Neural Network(ANN)model is developed to predict the thermal conductivity of metal oxide water-based nanofluid.For this,a comprehensive set of 691 data points was collected from the literature.This dataset is split into training(70%),validation(15%),and testing(15%)and used to train the ANN model.The developed model is a backpropagation artificial neural network with a 4–12–1 architecture.The performance of the developed model shows high accuracy with R values above 0.90 and rapid convergence.It shows that the developed ANN model accurately predicts the thermal conductivity of nanofluids.
基金co-supported by the National Science and Technology Major Project(No.J2019-Ⅲ-0010-0054)the National Natural Science Foundation of China(No.52336002)。
文摘The flight envelope of Air Turbo Rocket(ATR)engines is broader compared to conventional aero-engines,and designing a full-envelope controller using traditional methods poses significant challenges due to a burdensome design process.To address this issue,this paper proposes a self-learning neural network controller design method based on Reinforcement Learning(RL).Additionally,a method for predictive compensation and stability rewards is proposed to reduce the system oscillation caused by actuator delay.This approach simplifies the actuator to a firstorder inertial element exhibiting pure delay.A simulation environment for the ATR engineactuator system is first established.Based on this environment,a self-learning neural network controller using a predictive compensator and the Proximal Policy Optimization(PPO)algorithm is then developed.Furthermore,the temporal difference signals from the controller output are integrated into the reward function to enhance system stability.The proposed method is validated through numerical simulations and semi-physical experiments.The numerical simulation results demonstrate that the proposed method increases the system's tolerance to delays from 20 ms to 400 ms.Under an actuator delay of 400 ms,the average steady-state error remains less than0.1%,the overshoot is limited to 1%,and the settling time does not exceed 3 s.Moreover,compared to the traditional method,the proposed method exhibits higher adaptability to model errors and variations in flight conditions.In the conducted semi-physical simulation experiments,the proposed method achieves stable control of a real electric pump.
基金supported by the Shanghai Pilot Program for Basic Research(22T01400100-18)the National Natural Science Foundation of China(22278127 and 12447149)+1 种基金the Fundamental Research Funds for the Central Universities(2022ZFJH004)the Postdoctoral Fellowship Program of CPSF(GZB20250159).
文摘Accurate state of health(SOH)estimation is essential for the safe and reliable operation of lithium-ion batteries.However,existing methods face significant challenges,primarily because they rely on complete charge–discharge cycles and fixed-form physical constraints,which limit adaptability to different chemistries and real-world conditions.To address these issues,this study proposes an approach that extracts features from segmented state of charge(SOC)intervals and integrates them into an enhanced physics-informed neural network(PINN).Specifically,voltage data within the 25%–75%SOC range during charging are used to derive statistical,time–frequency,and mechanism-based features that capture degradation trends.A hybrid PINN-Lasso-Transformer-BiLSTM architecture is developed,where Lasso regression enables sparse feature selection,and a nonlinear empirical degradation model is embedded as a learnable physical term within a dynamically scaled composite loss.This design adaptively balances data-driven accuracy with physical consistency,thereby enhancing estimation precision,robustness,and generalization.The results show that the proposed method outperforms conventional neural networks across four battery chemistries,achieving root mean square error and mean absolute error below 1%.Notably,features from partial charging segments exhibit higher robustness than those from full cycles.Furthermore,the model maintains strong performance under high temperatures and demonstrates excellent generalization capacity in transfer learning across chemistries,temperatures,and C-rates.This work establishes a scalable and interpretable solution for accurate SOH estimation under diverse practical operating conditions.
基金supported by the National Natural Science Foundation of China(Grant No.12272316).
文摘The adjoint method is widely used in gradient-based optimization with high-dimensional design variables.However,the cost of solving the adjoint equations in each iteration is comparable to that of solving the flow field,resulting in expensive computational costs.To improve the efficiency of solving adjoint equations,we propose a physics-constrained graph neural networks for solving adjoint equations,named ADJ-PCGN.ADJ-PCGN establishes a mapping relationship between flow characteristics and adjoint vector based on data,serving as a replacement for the computationally expensive numerical solution of adjoint equations.A physics-based graph structure and message-passing mechanism are designed to endow its strong fitting and generalization capabilities.Taking transonic drag reduction and maximum lift-drag ratio of the airfoil as examples,results indicate that ADJ-PCGN attains a similar optimal shape as the classical direct adjoint loop method.In addition,ADJ-PCGN demonstrates strong generalization capabilities across different mesh topologies,mesh densities,and out-of-distribution conditions.It holds the potential to become a universal model for aerodynamic shape optimization involving states,geometries,and meshes.
基金founded by the Ministry of Science and Higher Education of the Russian Federation,State assignments for research,registration No.1024032600084-8-1.3.2Study of the grain growth and the formation of polycrystalline structure as a result of phase transition(Section 6)was founded by the Russian Science Foundation,Project No.24-71-00078+3 种基金https://rscf.ru/en/project/24-71-00078/(accessed on 01 December 2025).Study of the orientation dependence of the phase transition of aluminum in Section 3 was founded by the Russian Science Foundation,Project No.24-19-00684https://rscf.ru/en/project/24-19-00684/(accessed on 01 December 2025).
文摘It is well known that aluminum and copper exhibit structural phase transformations in quasi-static and dynamic measurements,including shock wave loading.However,the dependence of phase transformations in a wide range of crystallographic directions of shock loading has not been revealed.In this work,we calculated the shock Hugoniot for aluminum and copper in different crystallographic directions([100],[110],[111],[112],[102],[114],[123],[134],[221]and[401])of shock compression using molecular dynamics(MD)simulations.The results showed a high pressure(>160 GPa for Cu and>40 GPa for Al)of the FCC-to-BCC transition.In copper,different characteristics of the phase transition are observed depending on the loading direction with the[100]compression direction being the weakest.The FCC-to-BCC transition for copper is in the range of 150–220 GPa,which is consistent with the existing experimental data.Due to the high transition pressure,the BCC phase transition in copper competes with melting.In aluminum,the FCC-to-BCC transition is observed for all studied directions at pressures between 40 and 50 GPa far beyond the melting.In all considered cases we observe the coexistence of HCP and BCC phases during the FCC-to-BCC transition,which is consistent with the experimental data and atomistic calculations;this HCP phase forms in the course of accompanying plastic deformation with dislocation activity in the parent FCC phase.The plasticity incipience is also anisotropic in bothmetals,which is due to the difference in the projections of stress on the slip plane for different orientations of the FCC crystal.MD modeling results demonstrate a strong dependence of the FCC-to-BCC transition on the crystallographic direction,in which the material is loaded in the copper crystals.However,MD simulations data can only be obtained for specific points in the stereographic direction space;therefore,for more comprehensive understanding of the phase transition process,a feed-forward neural network was trained using MD modeling data.The trained machine learning model allowed us to construct continuous stereographic maps of phase transitions as a function of stress in the shock-compressed state of metal.Due to appearance and growth of multiple centers of new phase,the FCC-to-BCC transition leads to formation of a polycrystalline structure from the parent single crystal.
基金supported by the National Science and Technology Major Project(2022ZD0119901)the National Natural Science Foundation of China under Grant(U2141234,62463004 and U24A20260)+1 种基金the Hainan Province Science and Technology Special Fund(ZDYF2024GXJS003)the Scientific Research Fund of Hainan University(KYQD(ZR)23025).
文摘Recently,the zeroing neural network(ZNN)has demonstrated remarkable effectiveness in tackling time-varying problems,delivering robust performance across both noise-free and noisy environments.However,existing ZNN models are limited in their ability to actively suppress noise,which constrains their robustness and precision in solving time-varying problems.This paper introduces a novel active noise rejection ZNN(ANR-ZNN)design that enhances noise suppression by integrating computational error dynamics and harmonic behaviour.Through rigorous theoretical analysis,we demonstrate that the proposed ANR-ZNN maintains robust convergence in computational error performance under environmental noise.As a case study,the ANR-ZNN model is specifically applied to time-varying matrix inversion.Comprehensive computer simulations and robotic experiments further validate the ANR-ZNN's effectiveness,emphasising the proposed design's superiority and potential for solving time-varying problems.
基金supported by the Korea Institute of Energy Technology Evaluation and Planning(KETEP)and the Ministry of Trade,Industry&Energy(MOTIE)of the Republic of Korea(No.RS-2025-02315209).
文摘There is a need for accurate prediction of heat and mass transfer in aerodynamically designed,non-Newtonian nanofluids across aerodynamically designed,high-flux biomedical micro-devices for thermal management and reactive coating processes,but existing work is not uncharacteristically remiss regarding viscoelasticity,radiative heating,viscous dissipation,and homogeneous–heterogeneous reactions within a single scheme that is calibrated.This research investigates the flow of Williamson nanofluid across a dynamically wedged surface under conditions that include viscous dissipation,thermal radiation,and homogeneous-heterogeneous reactions.The paper develops a detailed mathematical approach that utilizes boundary layers to transform partial differential equations into ordinary differential equations using similarity transformations.RK4 is the technique for gaining numerical solutions,but with the addition of ANNs,there is an improvement in prediction accuracy and computational efficiency.The study investigates the influence of wedge angle parameter,along with Weissenberg number,thermal radiation parameter and Brownian motion parameter,and Schmidt number,on velocity distribution,temperature distribution,and concentra-tion distribution.Enhanced Weissenberg numbers enhance viscoelastic responses that modify velocity patterns,but radiation parameters and thermophoresis have key impacts on thermal transfer phenomena.This research develops findings that are of enormous application in aerospace,biomedical(artificial hearts and drug delivery),and industrial cooling technology applications.New findings on non-Newtonian nanofluids under full flow systems are included in this work to enhance heat transfer methods in novel fluid-based systems.
基金supported by the National Key Research and Development Program of China(2020YFB1005704).
文摘Since the advent of smart contracts,security vulnerabilities have remained a persistent challenge,compromsing both the reliability of contract execution and the overall stability of the virtual currency market.Consequently,the academic community has devoted increasing attention to these security risks.However,conventional approaches to vulnerability detection frequently exhibit limited accuracy.To address this limitation,the present study introduces a novel vulnerability detection framework called GNNSE that integrates symbolic execution with graph neural networks(GNNs).The proposedmethod first constructs semantic graphs to comprehensively capture the control flow and data flow dependencies within smart contracts.These graphs are subsequently processed using GNNs to efficiently identify contracts with a high likelihood of vulnerabilities.For these high-risk contracts,symbolic execution is employed to perform fine-grained,path-level analysis,thereby improving overall detection precision.Experimental results on a dataset comprising 10,079 contracts demonstrate that the proposed method achieves detection precisions of 93.58% for reentrancy vulnerabilities and 92.73% for timestamp-dependent vulnerabilities.
文摘In contemporary computer vision,convolutional neural networks(CNNs)and vision transformers(ViTs)represent the two primary architectural paradigms for image recognition.While both approaches have been widely adopted in medical imaging applications,they operate based on fundamentally different computational principles.This report attempts to provide brief application notes on ViTs and CNNs,particularly focusing on scenarios that guide the selection of one architecture over the other in practical medical implementations.Generally,CNNs rely on convolutional kernels,localized receptive fields,and weight sharing,enabling efficient hierarchical feature extraction.These properties contribute to strong performance in detecting spatially constrained patterns such as textures,edges,and anatomical boundaries,while maintaining relatively low computational requirements.ViTs,on the other hand,decompose images into smaller segments referred to as tokens and employ self-attention mechanisms to model relationships across the entire image.This global modeling capability allows ViTs to capture long-range dependencies that may be difficult for convolution-based architectures to learn.However,ViTs typically achieve optimal performance when trained on extremely large datasets or when supported by extensive pretraining,as their reduced inductive bias requires greater data exposure to learn robust representations.This report briefly examines the architectural structure,underlying mathematical foundations,and relative performance characteristics of CNNs and ViTs,drawing upon recent findings from contemporary research.Emphasis is placed on understanding how differences in data availability,computational resources,and task requirements influence model effectiveness across medical imaging domains.Most importantly,the report serves as a concise application guide for practitioners seeking informed implementation decisions between these two influential deep learning frameworks.
基金supported by the Guangdong S&T Program(Grant No.2025B1111130003).
文摘The development of β-titanium alloys with bone-mimicking elastic moduli remains a significant challenge.Although machine learning has the potential to accelerate alloy discovery,traditional methods often face data limitations such as sparsity,compositional discontinuity,and feature heterogeneity,leading to overfitting and restricting the exploration of novel compositional spaces.In this study,we introduce a domain-adversarial neural network framework that balances predictive accuracy with the generalization ability of unexplored composition space through integrated feature alignment and adversarial training.Using this approach,we successfully developed a non-intuitiveβ-Ti alloy with an ultra-low elastic modulus of 28±3 GPa,providing new insights beyond conventionally designed biomedical titanium alloys.This work establishes a screening framework for materials discovery in small-sample data spaces,with broad implications for the design of biomedical and other alloy systems.
