The soil-water retention curve(SWRC)plays a pivotal role in understanding water movement across numerous geological engineering applications.Despite significant advancements in theoretical modeling approaches,accurate...The soil-water retention curve(SWRC)plays a pivotal role in understanding water movement across numerous geological engineering applications.Despite significant advancements in theoretical modeling approaches,accurate prediction of SWRCs remains challenging due to the inherently sparse and incomplete nature of site-specific data.This study compiled a comprehensive dataset of SWRCs spanning a wide suction range from various published literature sources.Based on this dataset,multiple machine learning(ML)algorithms were employed to predict SWRCs.The performance of each algorithm was evaluated and ranked using four statistical indicators that quantify simulation accuracy.Feature importance analysis was subsequently conducted to reduce dimensionality by eliminating weakly correlated variables,thereby enhancing both model adaptability and computational efficiency.Following dimensionality reduction,a base learner pool was constructed and integrated through stacked generalization to create a multi-algorithm ensemble model.The proposed stacked model demonstrated robust performance in simulating SWRCs across diverse soil types,using only basic physical properties as inputs,achieving accuracy comparable to or marginally superior to the LightGBM model.The principal advantage of the stacked approach lies in its substantially improved accuracy within high suction ranges,effectively overcoming the limitations observed in LightGBM and enhancing the estimation under these conditions.This study provides valuable insights for researchers evaluating SWRCs through ML algorithms and demonstrates the potential of ensemble techniques in geotechnical prediction tasks.展开更多
The overgeneralization of fear is associated with psychiatric disorders and cognitive decline.Recent studies have shown that engram cells in the dorsal dentate gyrus are integrated into functionally heterogeneous ense...The overgeneralization of fear is associated with psychiatric disorders and cognitive decline.Recent studies have shown that engram cells in the dorsal dentate gyrus are integrated into functionally heterogeneous ensembles that are involved in contextual fear memory generalization and discrimination.However,the intracellular signals that promote fear generalization remain to be fully elucidated.In this study,we labeled and manipulated the c-Fos+and Npas4+ensembles in the dorsal dentate gyrus that are activated by contextual fear conditioning using a robust activity marking system.The results showed that increasing the excitability of Fos-dependent robust activity marking by overexpressing NaChBac or decreasing the excitability of Npas4-dependent robust activity marking by overexpressing Kir2.1 promoted fear memory generalization.Furthermore,CRISPR-mediated downregulation of the autophagy-related Atg5 or Atg7 genes in dorsal dentate gyrus neurons inhibited activation of c-Fos,but not Npas4.Knockdown of Atg5 or Atg7 in the Fos-dependent robust activity marking or Npas4-dependent robust activity marking ensemble led to an increase in neuronal excitability and a decrease in spine density in both ensembles.However,Atg7 knockdown in the Fos-dependent robust activity marking ensemble promoted memory generalization,while knockdown of Atg5 or Atg7 in the Npas4-dependent robust activity marking ensemble increased anxiety levels.These results contribute to our understanding of how the varying plasticity of memory engrams is involved in regulating fear memory generalization and anxiety.展开更多
Face liveness detection is essential for securing biometric authentication systems against spoofing attacks,including printed photos,replay videos,and 3D masks.This study systematically evaluates pre-trained CNN model...Face liveness detection is essential for securing biometric authentication systems against spoofing attacks,including printed photos,replay videos,and 3D masks.This study systematically evaluates pre-trained CNN models—DenseNet201,VGG16,InceptionV3,ResNet50,VGG19,MobileNetV2,Xception,and InceptionResNetV2—leveraging transfer learning and fine-tuning to enhance liveness detection performance.The models were trained and tested on NUAA and Replay-Attack datasets,with cross-dataset generalization validated on SiW-MV2 to assess real-world adaptability.Performance was evaluated using accuracy,precision,recall,FAR,FRR,HTER,and specialized spoof detection metrics(APCER,NPCER,ACER).Fine-tuning significantly improved detection accuracy,with DenseNet201 achieving the highest performance(98.5%on NUAA,97.71%on Replay-Attack),while MobileNetV2 proved the most efficient model for real-time applications(latency:15 ms,memory usage:45 MB,energy consumption:30 mJ).A statistical significance analysis(paired t-tests,confidence intervals)validated these improvements.Cross-dataset experiments identified DenseNet201 and MobileNetV2 as the most generalizable architectures,with DenseNet201 achieving 86.4%accuracy on Replay-Attack when trained on NUAA,demonstrating robust feature extraction and adaptability.In contrast,ResNet50 showed lower generalization capabilities,struggling with dataset variability and complex spoofing attacks.These findings suggest that MobileNetV2 is well-suited for low-power applications,while DenseNet201 is ideal for high-security environments requiring superior accuracy.This research provides a framework for improving real-time face liveness detection,enhancing biometric security,and guiding future advancements in AI-driven anti-spoofing techniques.展开更多
The phenomenon of fear memory generalization can be defined as the expansion of an individual's originally specific fear responses to a similar yet genuinely harmless stimulus or situation subsequent to the occurr...The phenomenon of fear memory generalization can be defined as the expansion of an individual's originally specific fear responses to a similar yet genuinely harmless stimulus or situation subsequent to the occurrence of a traumatic event[1].Fear generalization within the normal range represents an adaptive evolutionary mechanism to facilitate prompt reactions to potential threats and to enhance the likelihood of survival.展开更多
The challenge of enhancing the generalization capacity of reinforcement learning(RL)agents remains a formidable obstacle.Existing RL methods,despite achieving superhuman performance on certain benchmarks,often struggl...The challenge of enhancing the generalization capacity of reinforcement learning(RL)agents remains a formidable obstacle.Existing RL methods,despite achieving superhuman performance on certain benchmarks,often struggle with this aspect.A potential reason is that the benchmarks used for training and evaluation may not adequately offer a diverse set of transferable tasks.Although recent studies have developed bench-marking environments to address this shortcoming,they typically fall short in providing tasks that both ensure a solid foundation for generalization and exhibit significant variability.To overcome these limitations,this work introduces the concept that‘objects are composed of more fundamental components’in environment design,as implemented in the proposed environment called summon the magic(StM).This environment generates tasks where objects are derived from extensible and shareable basic components,facilitating strategy reuse and enhancing generalization.Furthermore,two new metrics,adaptation sensitivity range(ASR)and parameter correlation coefficient(PCC),are proposed to better capture and evaluate the generalization process of RL agents.Experimental results show that increasing the number of basic components of the object reduces the proximal policy optimization(PPO)agent’s training-testing gap by 60.9%(in episode reward),significantly alleviating overfitting.Additionally,linear variations in other environmental factors,such as the training monster set proportion and the total number of basic components,uniformly decrease the gap by at least 32.1%.These results highlight StM’s effectiveness in benchmarking and probing the generalization capabilities of RL algorithms.展开更多
Using in-situ microstructure observations from 2010 to 2018,this study investigates the performance and generalization of machine learning models in parameterizing turbulent mixing in the northwestern South China Sea....Using in-situ microstructure observations from 2010 to 2018,this study investigates the performance and generalization of machine learning models in parameterizing turbulent mixing in the northwestern South China Sea.The results show that the data-driven extreme gradient boosting(XGBoost)performs better than the other four models,i.e.,random forest,neural network,linear regression and support vector machine regression.In order to further improve the generalization of machine learning-based parameterization method,we propose a physics-informed machine learning(PIML)that couples the MacKinnon-Gregg model(known as the MG model)and Osborn’s formula to the XGBoost model.The correlation coefficient(r)and root mean square error(RMSE)between the estimated and observed 1g(ε)(whereεdenotes the turbulent kinetic energy dissipation rate)from the PIML are improved by 14%and 16%,respectively.The results also show that PIML effectively improves the generalization of the XGBoost-based parameterization method,enhancing r and RMSE by 35%and 75%,respectively.展开更多
Automatically recognizing radar emitters from com-plex electromagnetic environments is important but non-trivial.Moreover,the changing electromagnetic environment results in inconsistent signal distribution in the rea...Automatically recognizing radar emitters from com-plex electromagnetic environments is important but non-trivial.Moreover,the changing electromagnetic environment results in inconsistent signal distribution in the real world,which makes the existing approaches perform poorly for recognition tasks in different scenes.In this paper,we propose a domain generaliza-tion framework is proposed to improve the adaptability of radar emitter signal recognition in changing environments.Specifically,we propose an end-to-end denoising based domain-invariant radar emitter recognition network(DDIRNet)consisting of a denoising model and a domain invariant representation learning model(IRLM),which mutually benefit from each other.For the signal denoising model,a loss function is proposed to match the feature of the radar signals and guarantee the effectiveness of the model.For the domain invariant representation learning model,contrastive learning is introduced to learn the cross-domain feature by aligning the source and unseen domain distri-bution.Moreover,we design a data augmentation method that improves the diversity of signal data for training.Extensive experiments on classification have shown that DDIRNet achieves up to 6.4%improvement compared with the state-of-the-art radar emitter recognition methods.The proposed method pro-vides a promising direction to solve the radar emitter signal recognition problem.展开更多
In the task of domain generalization person re-identification(ReID),pedestrian image features exhibit significant intraclass variability and inter-class similarity.Existing methods rely on a single feature extraction ...In the task of domain generalization person re-identification(ReID),pedestrian image features exhibit significant intraclass variability and inter-class similarity.Existing methods rely on a single feature extraction architecture and struggle to capture both global context and local spatial information,resulting in weaker generalization to unseen domains.To address this issue,an innovative domain generalization person ReID method–LViT-Net,which combines local semantics and multi-feature cross fusion,is proposed.LViT-Net adopts a dual-branch encoder with a parallel hierarchical structure to extract both local and global discriminative features.In the local branch,the local multi-scale feature fusion module is designed to fuse local feature units at different scales to ensure that the fine-grained local features at various levels are accurately captured,thereby enhancing the robustness of the features.In the global branch,the dual feature cross fusion module fuses local features and global semantic information,focusing on critical semantic information and enabling the mutual refinement and matching of local and global features.This allows the model to achieve a dynamic balance between detailed and holistic information,forming robust feature representations of pedestrians.Extensive experiments demonstrate the effectiveness of LViT-Net.In both single-source and multisource comparison experiments,the proposed method outperforms existing state-of-the-art methods.展开更多
In actual industrial scenarios,the variation of operating conditions,the existence of data noise,and failure of measurement equipment will inevitably affect the distribution of perceptive data.Deep learning-based faul...In actual industrial scenarios,the variation of operating conditions,the existence of data noise,and failure of measurement equipment will inevitably affect the distribution of perceptive data.Deep learning-based fault diagnosis algorithms strongly rely on the assumption that source and target data are independent and identically distributed,and the learned diagnosis knowledge is difficult to generalize to out-of-distribution data.Domain generalization(DG)aims to achieve the generalization of arbitrary target domain data by using only limited source domain data for diagnosis model training.The research of DG for fault diagnosis has made remarkable progress in recent years and lots of achievements have been obtained.In this article,for the first time a comprehensive literature review on DG for fault diagnosis from a learning mechanism-oriented perspective is provided to summarize the development in recent years.Specifically,we first conduct a comprehensive review on existing methods based on the similarity of basic principles and design motivations.Then,the recent trend of DG for fault diagnosis is also analyzed.Finally,the existing problems and future prospect is performed.展开更多
In the realm of medical image segmentation,particularly in cardiac magnetic resonance imaging(MRI),achieving robust performance with limited annotated data is a significant challenge.Performance often degrades when fa...In the realm of medical image segmentation,particularly in cardiac magnetic resonance imaging(MRI),achieving robust performance with limited annotated data is a significant challenge.Performance often degrades when faced with testing scenarios from unknown domains.To address this problem,this paper proposes a novel semi-supervised approach for cardiac magnetic resonance image segmentation,aiming to enhance predictive capabilities and domain generalization(DG).This paper establishes an MT-like model utilizing pseudo-labeling and consistency regularization from semi-supervised learning,and integrates uncertainty estimation to improve the accuracy of pseudo-labels.Additionally,to tackle the challenge of domain generalization,a data manipulation strategy is introduced,extracting spatial and content-related information from images across different domains,enriching the dataset with a multi-domain perspective.This papers method is meticulously evaluated on the publicly available cardiac magnetic resonance imaging dataset M&Ms,validating its effectiveness.Comparative analyses against various methods highlight the out-standing performance of this papers approach,demonstrating its capability to segment cardiac magnetic resonance images in previously unseen domains even with limited annotated data.展开更多
This paper analyzes the generalization of minimax regret optimization(MRO)under distribution shift.A new learning framework is proposed by injecting the measure of con-ditional value at risk(CVaR)into MRO,and its gene...This paper analyzes the generalization of minimax regret optimization(MRO)under distribution shift.A new learning framework is proposed by injecting the measure of con-ditional value at risk(CVaR)into MRO,and its generalization error bound is established through the lens of uniform convergence analysis.The CVaR-based MRO can achieve the polynomial decay rate on the excess risk,which extends the generalization analysis associated with the expected risk to the risk-averse case.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
In mixture experiments,the observed response is determined by the relative proportions of the components,consequently rendering the experimental region a simplex.This paper focuses primarily on the optimal designs of ...In mixture experiments,the observed response is determined by the relative proportions of the components,consequently rendering the experimental region a simplex.This paper focuses primarily on the optimal designs of mixture experiments that involve process variables.Prior research has extensively delved into optimal orthogonal block designs for some classic mixture models with process variables.Based on the framework of general blending models,this paper proposes a class of symmetric linear mixture models,which can be regarded as a generalization of many existing ones.Under the orthogonal blocking conditions,orthogonal block designs are devised through Latin squares in the presence of process variables.TheD-,A-,and E-optimality criteria are utilized to obtain optimal designs at the boundary of the simplex in the case of 3 components.As the values of the exponents change,numerically derived optimal design points are presented to illustrate the pattern of their variations,and to verify the consistency of the results with previous research on some specific symmetric general blending models.展开更多
Clouds play an important role in global atmospheric energy and water vapor budgets, and the low cloud simulations suffer from large biases in many atmospheric general circulation models. In this study, cloud microphys...Clouds play an important role in global atmospheric energy and water vapor budgets, and the low cloud simulations suffer from large biases in many atmospheric general circulation models. In this study, cloud microphysical processes such as raindrop evaporation and cloud water accretion in a double-moment six-class cloud microphysics scheme were revised to enhance the simulation of low clouds using the Global-Regional Integrated Forecast System(GRIST)model. The validation of the revised scheme using a single-column version of the GRIST demonstrated a reasonable reduction in liquid water biases. The revised parameterization simulated medium-and low-level cloud fractions that were in better agreement with the observations than the original scheme. Long-term global simulations indicate the mitigation of the originally overestimated low-level cloud fraction and cloud-water mixing ratio in mid-to high-latitude regions,primarily owing to enhanced accretion processes and weakened raindrop evaporation. The reduced low clouds with the revised scheme showed better consistency with satellite observations, particularly at mid-and high-latitudes. Further improvements can be observed in the simulated cloud shortwave radiative forcing and vertical distribution of total cloud cover. Annual precipitation in mid-latitude regions has also improved, particularly over the oceans, with significantly increased large-scale and decreased convective precipitation.展开更多
The construction of new agricultural science has put forward the core requirements of"interdisciplinary integration,service industry demand,and cultivation of composite talents"for the smart agriculture majo...The construction of new agricultural science has put forward the core requirements of"interdisciplinary integration,service industry demand,and cultivation of composite talents"for the smart agriculture major.The"integration of general and specialized education"is the key path to solve the problems of"prominent disciplinary barriers,fragmented knowledge structure,and weak practical ability"in the traditional curriculum system.In this paper,the College of Smart Agriculture from Yulin Normal University is taken as the research object.Based on the characteristics of regional agricultural industry and the positioning of professional education,the prominent problems in the current professional curriculum system of smart agriculture are analyzed,the construction concept of"strong foundation in general education,precise core in professional education,and breaking through boundaries in integrated education"is proposed,and a"three dimensions and four layers"integrated curriculum system framework for general and specialized education is constructed.Moreover,practical exploration is conducted from the aspects of curriculum module design,teaching mode innovation,and guarantee mechanism construction.Practice has shown that this curriculum system effectively enhances students'interdisciplinary application abilities and industry adaptability,and provides a practical sample for the reform of smart agriculture courses in local universities under the background of new agricultural science.展开更多
Making full use of the operator ordering method and the integration within ordered products,we obtain the analytical evolution law of a general quadratic state in the amplitude decay channel,and find that it is determ...Making full use of the operator ordering method and the integration within ordered products,we obtain the analytical evolution law of a general quadratic state in the amplitude decay channel,and find that it is determined not only by the decay rate of the amplitude decay channel but also by the coefficients of the initial quadratic state.Further,the quantum statistical properties of the initial quadratic state for amplitude decay are investigated via its average photon number and photon-counting distribution,and its Wigner distribution function evolution is discussed in detail.展开更多
Modern apprenticeship emphasizes strengthening school-enterprise cooperation,deepening the integration of production and education,and promoting the combination of work and study to achieve seamless connection between...Modern apprenticeship emphasizes strengthening school-enterprise cooperation,deepening the integration of production and education,and promoting the combination of work and study to achieve seamless connection between professional education and industrial needs,thereby improving the quality of talent training.This paper analyzes the current problems in the nursing talent training system,elaborates on the significance of cultivating nursing talents based on modern apprenticeship,and proposes new ideas for constructing a modern apprenticeship training model for nursing talents from the perspectives of“integrating general and specialized education,merging virtue and skills,and linking specialization and innovation”.The aim is to further promote the organic combination of nursing talent training goals and post needs,for reference only.展开更多
This paper puts forword 11 cartographic generalization operator models and introduces their mathematical definitions,and thus a precise mathematical form and quantitative description has been given to these formerly l...This paper puts forword 11 cartographic generalization operator models and introduces their mathematical definitions,and thus a precise mathematical form and quantitative description has been given to these formerly limited qualitative concepts.The meaning of mathematical definition of operators for cartographic generalization and the application prospect in computer_aided cartography (CAC) is stated.ract The Jurassic strata in Jingyan of Sichuan containing the Mamenchinsaurus fauna are dealt with and divided in this paper. The Mamenchisaurus fossils contained there are compared in morphological features and stratigraphically with other types of the genus on by one. The comprehensive analysis show that the Mamenchisaurus fauna of Jingyan appeared in the early Late Jurassic and is primitive in morphology. The results of the morphological identification and stratigraphical study agree with each other. Their evolutionary processes in different apoches of the Late Jurassic also made clear. Key words Jingyan, Sichuan, Mamenchisaurus Fauna, stratigraphy, evolution展开更多
Generalization ability is a major problem encountered when using neural networks to find the structures in noisy data sets. Controlling the network complexity is a common method to solve this problem. In this paper, h...Generalization ability is a major problem encountered when using neural networks to find the structures in noisy data sets. Controlling the network complexity is a common method to solve this problem. In this paper, however, a novel additive penalty term which represents the features extracted by hidden units is introduced to eliminate the overtraining of multilayer feedfoward networks. Computer simulations demonstrate that by using this unsupervised fashion penalty term, the generalization ability is greatly improved.展开更多
This paper studies the generalization capability of feedforward neural networks (FNN).The mechanism of FNNs for classification is investigated from the geometric and probabilistic viewpoints. It is pointed out that th...This paper studies the generalization capability of feedforward neural networks (FNN).The mechanism of FNNs for classification is investigated from the geometric and probabilistic viewpoints. It is pointed out that the outputs of the output layer in the FNNs for classification correspond to the estimates of posteriori probability of the input pattern samples with desired outputs 1 or 0. The theorem for the generalized kernel function in the radial basis function networks (RBFN) is given. For an 2-layer perceptron network (2-LPN). an idea of using extended samples to improve generalization capability is proposed. Finally. the experimental results of radar target classification are given to verify the generaliztion capability of the RBFNs.展开更多
基金the National Natural Science Foundation of China(Grant No.42272312)Ningbo Youth Science and Technology Innovation Talent Project(Grant No.2024QL057)the Zhejiang Provincial Xinmiao Talents Program(Grant No.2024R405B093).
文摘The soil-water retention curve(SWRC)plays a pivotal role in understanding water movement across numerous geological engineering applications.Despite significant advancements in theoretical modeling approaches,accurate prediction of SWRCs remains challenging due to the inherently sparse and incomplete nature of site-specific data.This study compiled a comprehensive dataset of SWRCs spanning a wide suction range from various published literature sources.Based on this dataset,multiple machine learning(ML)algorithms were employed to predict SWRCs.The performance of each algorithm was evaluated and ranked using four statistical indicators that quantify simulation accuracy.Feature importance analysis was subsequently conducted to reduce dimensionality by eliminating weakly correlated variables,thereby enhancing both model adaptability and computational efficiency.Following dimensionality reduction,a base learner pool was constructed and integrated through stacked generalization to create a multi-algorithm ensemble model.The proposed stacked model demonstrated robust performance in simulating SWRCs across diverse soil types,using only basic physical properties as inputs,achieving accuracy comparable to or marginally superior to the LightGBM model.The principal advantage of the stacked approach lies in its substantially improved accuracy within high suction ranges,effectively overcoming the limitations observed in LightGBM and enhancing the estimation under these conditions.This study provides valuable insights for researchers evaluating SWRCs through ML algorithms and demonstrates the potential of ensemble techniques in geotechnical prediction tasks.
基金the STI2030-Major Projects,Nos.2021ZD0203500(to FW),2021ZD0202100(to XL)the National Natural Science Foundation of China,Nos.32222033(to FW),32330041(to LM)and 82021002(to LM),32171041(to XL)and 32450102(to XL)CAMS Innovation Fund for Medical Sciences,No.2021-I2M-5-009(to LM and XL).
文摘The overgeneralization of fear is associated with psychiatric disorders and cognitive decline.Recent studies have shown that engram cells in the dorsal dentate gyrus are integrated into functionally heterogeneous ensembles that are involved in contextual fear memory generalization and discrimination.However,the intracellular signals that promote fear generalization remain to be fully elucidated.In this study,we labeled and manipulated the c-Fos+and Npas4+ensembles in the dorsal dentate gyrus that are activated by contextual fear conditioning using a robust activity marking system.The results showed that increasing the excitability of Fos-dependent robust activity marking by overexpressing NaChBac or decreasing the excitability of Npas4-dependent robust activity marking by overexpressing Kir2.1 promoted fear memory generalization.Furthermore,CRISPR-mediated downregulation of the autophagy-related Atg5 or Atg7 genes in dorsal dentate gyrus neurons inhibited activation of c-Fos,but not Npas4.Knockdown of Atg5 or Atg7 in the Fos-dependent robust activity marking or Npas4-dependent robust activity marking ensemble led to an increase in neuronal excitability and a decrease in spine density in both ensembles.However,Atg7 knockdown in the Fos-dependent robust activity marking ensemble promoted memory generalization,while knockdown of Atg5 or Atg7 in the Npas4-dependent robust activity marking ensemble increased anxiety levels.These results contribute to our understanding of how the varying plasticity of memory engrams is involved in regulating fear memory generalization and anxiety.
基金funded by Centre for Advanced Modelling and Geospatial Information Systems(CAMGIS),Faculty of Engineering and IT,University of Technology Sydney.Moreover,Ongoing Research Funding Program(ORF-2025-14)King Saud University,Riyadh,Saudi Arabia,under Project ORF-2025-。
文摘Face liveness detection is essential for securing biometric authentication systems against spoofing attacks,including printed photos,replay videos,and 3D masks.This study systematically evaluates pre-trained CNN models—DenseNet201,VGG16,InceptionV3,ResNet50,VGG19,MobileNetV2,Xception,and InceptionResNetV2—leveraging transfer learning and fine-tuning to enhance liveness detection performance.The models were trained and tested on NUAA and Replay-Attack datasets,with cross-dataset generalization validated on SiW-MV2 to assess real-world adaptability.Performance was evaluated using accuracy,precision,recall,FAR,FRR,HTER,and specialized spoof detection metrics(APCER,NPCER,ACER).Fine-tuning significantly improved detection accuracy,with DenseNet201 achieving the highest performance(98.5%on NUAA,97.71%on Replay-Attack),while MobileNetV2 proved the most efficient model for real-time applications(latency:15 ms,memory usage:45 MB,energy consumption:30 mJ).A statistical significance analysis(paired t-tests,confidence intervals)validated these improvements.Cross-dataset experiments identified DenseNet201 and MobileNetV2 as the most generalizable architectures,with DenseNet201 achieving 86.4%accuracy on Replay-Attack when trained on NUAA,demonstrating robust feature extraction and adaptability.In contrast,ResNet50 showed lower generalization capabilities,struggling with dataset variability and complex spoofing attacks.These findings suggest that MobileNetV2 is well-suited for low-power applications,while DenseNet201 is ideal for high-security environments requiring superior accuracy.This research provides a framework for improving real-time face liveness detection,enhancing biometric security,and guiding future advancements in AI-driven anti-spoofing techniques.
基金supported by the Shandong Provincial Natural Science Foundation(ZR2022QH144).
文摘The phenomenon of fear memory generalization can be defined as the expansion of an individual's originally specific fear responses to a similar yet genuinely harmless stimulus or situation subsequent to the occurrence of a traumatic event[1].Fear generalization within the normal range represents an adaptive evolutionary mechanism to facilitate prompt reactions to potential threats and to enhance the likelihood of survival.
基金Supported by the National Key R&D Program of China(No.2023YFB4502200)the National Natural Science Foundation of China(No.U22A2028,61925208,62222214,62341411,62102398,62102399,U20A20227,62302478,62302482,62302483,62302480,62302481)+2 种基金the Strategic Priority Research Program of the Chinese Academy of Sciences(No.XDB0660300,XDB0660301,XDB0660302)the Chinese Academy of Sciences Project for Young Scientists in Basic Research(No.YSBR-029)the Youth Innovation Promotion Association of Chinese Academy of Sciences and Xplore Prize.
文摘The challenge of enhancing the generalization capacity of reinforcement learning(RL)agents remains a formidable obstacle.Existing RL methods,despite achieving superhuman performance on certain benchmarks,often struggle with this aspect.A potential reason is that the benchmarks used for training and evaluation may not adequately offer a diverse set of transferable tasks.Although recent studies have developed bench-marking environments to address this shortcoming,they typically fall short in providing tasks that both ensure a solid foundation for generalization and exhibit significant variability.To overcome these limitations,this work introduces the concept that‘objects are composed of more fundamental components’in environment design,as implemented in the proposed environment called summon the magic(StM).This environment generates tasks where objects are derived from extensible and shareable basic components,facilitating strategy reuse and enhancing generalization.Furthermore,two new metrics,adaptation sensitivity range(ASR)and parameter correlation coefficient(PCC),are proposed to better capture and evaluate the generalization process of RL agents.Experimental results show that increasing the number of basic components of the object reduces the proximal policy optimization(PPO)agent’s training-testing gap by 60.9%(in episode reward),significantly alleviating overfitting.Additionally,linear variations in other environmental factors,such as the training monster set proportion and the total number of basic components,uniformly decrease the gap by at least 32.1%.These results highlight StM’s effectiveness in benchmarking and probing the generalization capabilities of RL algorithms.
基金The National Science and Technology Major Project under contract No.2024YFC2817003the National Natural Science Foundation of China under contract Nos 42276019 and 42249911the Guangdong Ordinary University Innovation Team Project under contract No.2023KCXTD015.
文摘Using in-situ microstructure observations from 2010 to 2018,this study investigates the performance and generalization of machine learning models in parameterizing turbulent mixing in the northwestern South China Sea.The results show that the data-driven extreme gradient boosting(XGBoost)performs better than the other four models,i.e.,random forest,neural network,linear regression and support vector machine regression.In order to further improve the generalization of machine learning-based parameterization method,we propose a physics-informed machine learning(PIML)that couples the MacKinnon-Gregg model(known as the MG model)and Osborn’s formula to the XGBoost model.The correlation coefficient(r)and root mean square error(RMSE)between the estimated and observed 1g(ε)(whereεdenotes the turbulent kinetic energy dissipation rate)from the PIML are improved by 14%and 16%,respectively.The results also show that PIML effectively improves the generalization of the XGBoost-based parameterization method,enhancing r and RMSE by 35%and 75%,respectively.
基金supported by the National Natural Science Foundation of China(62101575)the Research Project of NUDT(ZK22-57)the Self-directed Project of State Key Laboratory of High Performance Computing(202101-16).
文摘Automatically recognizing radar emitters from com-plex electromagnetic environments is important but non-trivial.Moreover,the changing electromagnetic environment results in inconsistent signal distribution in the real world,which makes the existing approaches perform poorly for recognition tasks in different scenes.In this paper,we propose a domain generaliza-tion framework is proposed to improve the adaptability of radar emitter signal recognition in changing environments.Specifically,we propose an end-to-end denoising based domain-invariant radar emitter recognition network(DDIRNet)consisting of a denoising model and a domain invariant representation learning model(IRLM),which mutually benefit from each other.For the signal denoising model,a loss function is proposed to match the feature of the radar signals and guarantee the effectiveness of the model.For the domain invariant representation learning model,contrastive learning is introduced to learn the cross-domain feature by aligning the source and unseen domain distri-bution.Moreover,we design a data augmentation method that improves the diversity of signal data for training.Extensive experiments on classification have shown that DDIRNet achieves up to 6.4%improvement compared with the state-of-the-art radar emitter recognition methods.The proposed method pro-vides a promising direction to solve the radar emitter signal recognition problem.
基金supported by National Key Research and Development Program of China,No.2022YFB3305803.
文摘In the task of domain generalization person re-identification(ReID),pedestrian image features exhibit significant intraclass variability and inter-class similarity.Existing methods rely on a single feature extraction architecture and struggle to capture both global context and local spatial information,resulting in weaker generalization to unseen domains.To address this issue,an innovative domain generalization person ReID method–LViT-Net,which combines local semantics and multi-feature cross fusion,is proposed.LViT-Net adopts a dual-branch encoder with a parallel hierarchical structure to extract both local and global discriminative features.In the local branch,the local multi-scale feature fusion module is designed to fuse local feature units at different scales to ensure that the fine-grained local features at various levels are accurately captured,thereby enhancing the robustness of the features.In the global branch,the dual feature cross fusion module fuses local features and global semantic information,focusing on critical semantic information and enabling the mutual refinement and matching of local and global features.This allows the model to achieve a dynamic balance between detailed and holistic information,forming robust feature representations of pedestrians.Extensive experiments demonstrate the effectiveness of LViT-Net.In both single-source and multisource comparison experiments,the proposed method outperforms existing state-of-the-art methods.
基金supported by the National Natural Science Foundation of China(62322315,61873237)the Zhejiang Provincial Natural Science Foundation of China(LR22F030003)+1 种基金supported by Research Grant Council of Hong Kong(11201023,11202224)Hong Kong Innovation and Technology Commission(InnoHK Project CIMDA).
文摘In actual industrial scenarios,the variation of operating conditions,the existence of data noise,and failure of measurement equipment will inevitably affect the distribution of perceptive data.Deep learning-based fault diagnosis algorithms strongly rely on the assumption that source and target data are independent and identically distributed,and the learned diagnosis knowledge is difficult to generalize to out-of-distribution data.Domain generalization(DG)aims to achieve the generalization of arbitrary target domain data by using only limited source domain data for diagnosis model training.The research of DG for fault diagnosis has made remarkable progress in recent years and lots of achievements have been obtained.In this article,for the first time a comprehensive literature review on DG for fault diagnosis from a learning mechanism-oriented perspective is provided to summarize the development in recent years.Specifically,we first conduct a comprehensive review on existing methods based on the similarity of basic principles and design motivations.Then,the recent trend of DG for fault diagnosis is also analyzed.Finally,the existing problems and future prospect is performed.
基金Supported by the National Natural Science Foundation of China(No.62001313)the Key Project of Liaoning Provincial Department of Science and Technology(No.2021JH2/10300134,2022JH1/10500004)。
文摘In the realm of medical image segmentation,particularly in cardiac magnetic resonance imaging(MRI),achieving robust performance with limited annotated data is a significant challenge.Performance often degrades when faced with testing scenarios from unknown domains.To address this problem,this paper proposes a novel semi-supervised approach for cardiac magnetic resonance image segmentation,aiming to enhance predictive capabilities and domain generalization(DG).This paper establishes an MT-like model utilizing pseudo-labeling and consistency regularization from semi-supervised learning,and integrates uncertainty estimation to improve the accuracy of pseudo-labels.Additionally,to tackle the challenge of domain generalization,a data manipulation strategy is introduced,extracting spatial and content-related information from images across different domains,enriching the dataset with a multi-domain perspective.This papers method is meticulously evaluated on the publicly available cardiac magnetic resonance imaging dataset M&Ms,validating its effectiveness.Comparative analyses against various methods highlight the out-standing performance of this papers approach,demonstrating its capability to segment cardiac magnetic resonance images in previously unseen domains even with limited annotated data.
基金Supported by Education Science Planning Project of Hubei Province(2020GB198)Natural Science Foundation of Hubei Province(2023AFB523).
文摘This paper analyzes the generalization of minimax regret optimization(MRO)under distribution shift.A new learning framework is proposed by injecting the measure of con-ditional value at risk(CVaR)into MRO,and its generalization error bound is established through the lens of uniform convergence analysis.The CVaR-based MRO can achieve the polynomial decay rate on the excess risk,which extends the generalization analysis associated with the expected risk to the risk-averse case.
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 numbers 12071329,12471246].
文摘In mixture experiments,the observed response is determined by the relative proportions of the components,consequently rendering the experimental region a simplex.This paper focuses primarily on the optimal designs of mixture experiments that involve process variables.Prior research has extensively delved into optimal orthogonal block designs for some classic mixture models with process variables.Based on the framework of general blending models,this paper proposes a class of symmetric linear mixture models,which can be regarded as a generalization of many existing ones.Under the orthogonal blocking conditions,orthogonal block designs are devised through Latin squares in the presence of process variables.TheD-,A-,and E-optimality criteria are utilized to obtain optimal designs at the boundary of the simplex in the case of 3 components.As the values of the exponents change,numerically derived optimal design points are presented to illustrate the pattern of their variations,and to verify the consistency of the results with previous research on some specific symmetric general blending models.
基金National Natural Science Foundation of China(42375153,42105153,42205157)Development of Science and Technology at Chinese Academy of Meteorological Sciences(2023KJ038)。
文摘Clouds play an important role in global atmospheric energy and water vapor budgets, and the low cloud simulations suffer from large biases in many atmospheric general circulation models. In this study, cloud microphysical processes such as raindrop evaporation and cloud water accretion in a double-moment six-class cloud microphysics scheme were revised to enhance the simulation of low clouds using the Global-Regional Integrated Forecast System(GRIST)model. The validation of the revised scheme using a single-column version of the GRIST demonstrated a reasonable reduction in liquid water biases. The revised parameterization simulated medium-and low-level cloud fractions that were in better agreement with the observations than the original scheme. Long-term global simulations indicate the mitigation of the originally overestimated low-level cloud fraction and cloud-water mixing ratio in mid-to high-latitude regions,primarily owing to enhanced accretion processes and weakened raindrop evaporation. The reduced low clouds with the revised scheme showed better consistency with satellite observations, particularly at mid-and high-latitudes. Further improvements can be observed in the simulated cloud shortwave radiative forcing and vertical distribution of total cloud cover. Annual precipitation in mid-latitude regions has also improved, particularly over the oceans, with significantly increased large-scale and decreased convective precipitation.
基金Supported by the Autonomous Region-level Research and Practice Projects for New Engineering,New Medicine,New Agriculture,and New Humanities of Guangxi Department of Education(XNK202409)the Undergraduate Teaching Reform Project of Guangxi Higher Education(2024JGB332+1 种基金2024JGA304)the Guangxi Degree and Graduate Education Reform Project(JGY2025382).
文摘The construction of new agricultural science has put forward the core requirements of"interdisciplinary integration,service industry demand,and cultivation of composite talents"for the smart agriculture major.The"integration of general and specialized education"is the key path to solve the problems of"prominent disciplinary barriers,fragmented knowledge structure,and weak practical ability"in the traditional curriculum system.In this paper,the College of Smart Agriculture from Yulin Normal University is taken as the research object.Based on the characteristics of regional agricultural industry and the positioning of professional education,the prominent problems in the current professional curriculum system of smart agriculture are analyzed,the construction concept of"strong foundation in general education,precise core in professional education,and breaking through boundaries in integrated education"is proposed,and a"three dimensions and four layers"integrated curriculum system framework for general and specialized education is constructed.Moreover,practical exploration is conducted from the aspects of curriculum module design,teaching mode innovation,and guarantee mechanism construction.Practice has shown that this curriculum system effectively enhances students'interdisciplinary application abilities and industry adaptability,and provides a practical sample for the reform of smart agriculture courses in local universities under the background of new agricultural science.
文摘Making full use of the operator ordering method and the integration within ordered products,we obtain the analytical evolution law of a general quadratic state in the amplitude decay channel,and find that it is determined not only by the decay rate of the amplitude decay channel but also by the coefficients of the initial quadratic state.Further,the quantum statistical properties of the initial quadratic state for amplitude decay are investigated via its average photon number and photon-counting distribution,and its Wigner distribution function evolution is discussed in detail.
基金2025 Shanghai Secondary Vocational School Curriculum and Teaching Reform Research Project(Project No.:2025K-03)。
文摘Modern apprenticeship emphasizes strengthening school-enterprise cooperation,deepening the integration of production and education,and promoting the combination of work and study to achieve seamless connection between professional education and industrial needs,thereby improving the quality of talent training.This paper analyzes the current problems in the nursing talent training system,elaborates on the significance of cultivating nursing talents based on modern apprenticeship,and proposes new ideas for constructing a modern apprenticeship training model for nursing talents from the perspectives of“integrating general and specialized education,merging virtue and skills,and linking specialization and innovation”.The aim is to further promote the organic combination of nursing talent training goals and post needs,for reference only.
文摘This paper puts forword 11 cartographic generalization operator models and introduces their mathematical definitions,and thus a precise mathematical form and quantitative description has been given to these formerly limited qualitative concepts.The meaning of mathematical definition of operators for cartographic generalization and the application prospect in computer_aided cartography (CAC) is stated.ract The Jurassic strata in Jingyan of Sichuan containing the Mamenchinsaurus fauna are dealt with and divided in this paper. The Mamenchisaurus fossils contained there are compared in morphological features and stratigraphically with other types of the genus on by one. The comprehensive analysis show that the Mamenchisaurus fauna of Jingyan appeared in the early Late Jurassic and is primitive in morphology. The results of the morphological identification and stratigraphical study agree with each other. Their evolutionary processes in different apoches of the Late Jurassic also made clear. Key words Jingyan, Sichuan, Mamenchisaurus Fauna, stratigraphy, evolution
文摘Generalization ability is a major problem encountered when using neural networks to find the structures in noisy data sets. Controlling the network complexity is a common method to solve this problem. In this paper, however, a novel additive penalty term which represents the features extracted by hidden units is introduced to eliminate the overtraining of multilayer feedfoward networks. Computer simulations demonstrate that by using this unsupervised fashion penalty term, the generalization ability is greatly improved.
文摘This paper studies the generalization capability of feedforward neural networks (FNN).The mechanism of FNNs for classification is investigated from the geometric and probabilistic viewpoints. It is pointed out that the outputs of the output layer in the FNNs for classification correspond to the estimates of posteriori probability of the input pattern samples with desired outputs 1 or 0. The theorem for the generalized kernel function in the radial basis function networks (RBFN) is given. For an 2-layer perceptron network (2-LPN). an idea of using extended samples to improve generalization capability is proposed. Finally. the experimental results of radar target classification are given to verify the generaliztion capability of the RBFNs.
