Federated Learning(FL)enables joint training over distributed devices without data exchange but is highly vulnerable to attacks by adversaries in the form of model poisoning and malicious update injection.This work pr...Federated Learning(FL)enables joint training over distributed devices without data exchange but is highly vulnerable to attacks by adversaries in the form of model poisoning and malicious update injection.This work proposes Secured-FL,a blockchain-based defensive framework that combines smart contract-based authentication,clustering-driven outlier elimination,and dynamic threshold adjustment to defend against adversarial attacks.The framework was implemented on a private Ethereum network with a Proof-of-Authority consensus algorithm to ensure tamper-resistant and auditable model updates.Large-scale simulation on the Cyber Data dataset,under up to 50%malicious client settings,demonstrates Secured-FL achieves 6%-12%higher accuracy,9%-15%lower latency,and approximately 14%less computational expense compared to the PPSS benchmark framework.Additional tests,including confusion matrices,ROC and Precision-Recall curves,and ablation tests,confirm the interpretability and robustness of the defense.Tests for scalability also show consistent performance up to 500 clients,affirming appropriateness to reasonably large deployments.These results make Secured-FL a feasible,adversarially resilient FL paradigm with promising potential for application in smart cities,medicine,and other mission-critical IoT deployments.展开更多
This review explores the use of agent-based modeling(ABM)within the framework of study human emotion and cognition in the context of its ability to simulate complex social interactions,adaptive changes,and evolutionar...This review explores the use of agent-based modeling(ABM)within the framework of study human emotion and cognition in the context of its ability to simulate complex social interactions,adaptive changes,and evolutionary processes.By representing agents and their defined environments with probabilistic interactions,ABM allows the assessment of the effects of individual behavior at the micro level on the greater social phenomena at the macro level.The review looks into the applications of ABM in portraying some of the key components of emotions and cognition-empathy,cooperation,decision making,and emotional transmission-and analyzes the problems including scalability,empirical validation,and description of sensitive emotional states.The most important conclusion is that merging ABM with information neurobiological data and artificial intelligence(AI)techniques would allow for deepening the interactions within the system and enhancing its responsiveness to stimuli.This review highlights approaches that aim to exploit the ABM methodology more fully and integrates methods from biology,neuroscience,and engineering.This integration could contribute to our understanding of the human behavior evolution and adaptation within systems relevant to policymaking,healthcare,and education.展开更多
Conversational recommender systems(CRSs)focus on refining preferences and providing personalized recommendations through natural language interactions and dialogue history.Large language models(LLMs)have shown outstan...Conversational recommender systems(CRSs)focus on refining preferences and providing personalized recommendations through natural language interactions and dialogue history.Large language models(LLMs)have shown outstanding performance across various domains,thereby prompting researchers to investigate their applicability in recommendation systems.However,due to the lack of task-specific knowledge and an inefficient feature extraction process,LLMs still have suboptimal performance in recommendation tasks.Therefore,external knowledge sources,such as knowledge graphs(KGs)and knowledge bases(KBs),are often introduced to address the issue of data sparsity.Compared to KGs,KBs possess higher retrieval efficiency,making them more suitable for scenarios where LLMs serve as recommenders.To this end,we introduce a novel framework integrating LLMs with KBs for enhanced retrieval generation,namely LLMKB.LLMKB initially leverages structured knowledge to create mapping dictionaries,extracting entity-relation information from heterogeneous knowledge to construct KBs.Then,LLMKB achieves the embedding calibration between user information representations and documents in KBs through retrieval model fine-tuning.Finally,LLMKB employs retrievalaugmented generation to produce recommendations based on fused text inputs,followed by post-processing.Experiment results on two public CRS datasets demonstrate the effectiveness of our framework.Our code is publicly available at the link:https://anonymous.4open.science/r/LLMKB-6FD0.展开更多
Tau plays a crucial role in several neurodegenerative diseases,collectively referred to as tauopathies.Therefore,targeting potential pathological changes in tau could enable useful therapeutic interventions.However,ta...Tau plays a crucial role in several neurodegenerative diseases,collectively referred to as tauopathies.Therefore,targeting potential pathological changes in tau could enable useful therapeutic interventions.However,tau is not an easy target because it dynamically interacts with microtubules and other cellular components,which presents a challenge for tau-targeted drugs.New cellular models could aid the development of mechanism-based tau-targeted therapies.展开更多
Dear Editor,This letter presents a novel approach to the data-driven control of unknown nonlinear systems.By leveraging online sparse identification based on the Koopman operator,a high-dimensional linear system model...Dear Editor,This letter presents a novel approach to the data-driven control of unknown nonlinear systems.By leveraging online sparse identification based on the Koopman operator,a high-dimensional linear system model approximating the actual system is obtained online.The upper bound of the discrepancy between the identified model and the actual system is estimated using real-time prediction error,which is then utilized in the design of a tube-based robust model predictive controller.The effectiveness of the proposed approach is validated by numerical simulation.展开更多
Blend based polymer nanocomposites, comprising Janus nanoparticles at their polymer/polymer interface, were analytically/experimentally evaluated. The modeling procedure was performed in two stages: first, modeling of...Blend based polymer nanocomposites, comprising Janus nanoparticles at their polymer/polymer interface, were analytically/experimentally evaluated. The modeling procedure was performed in two stages: first, modeling of polymer/polymer interface region comprising Janus nanoparticles and second, modeling of the entire systems as a function of the variation of the blend morphology.In the first stage, the modeling procedure was performed based on the development of the model proposed by Ji et al. and in the second stage, the fundamental of Kolarik's model was used in order to propose a developed and more practical model. It was shown that Janus nanoparticles may form dual polymer/particle interphase at polymer/polymer interface which can drastically affect the final mechanical properties of the system. Comparing the results of tensile tests imposed on different prepared samples with the predictions of the model proved its accuracy and reliability(error < 9%).展开更多
Offline policy evaluation,evaluating and selecting complex policies for decision-making by only using offline datasets is important in reinforcement learning.At present,the model-based offline policy evaluation(MBOPE)...Offline policy evaluation,evaluating and selecting complex policies for decision-making by only using offline datasets is important in reinforcement learning.At present,the model-based offline policy evaluation(MBOPE)is widely welcomed because of its easy to implement and good performance.MBOPE directly approximates the unknown value of a given policy using the Monte Carlo method given the estimated transition and reward functions of the environment.Usually,multiple models are trained,and then one of them is selected to be used.However,a challenge remains in selecting an appropriate model from those trained for further use.The authors first analyse the upper bound of the difference between the approximated value and the unknown true value.Theoretical results show that this difference is related to the trajectories generated by the given policy on the learnt model and the prediction error of the transition and reward functions at these generated data points.Based on the theoretical results,a new criterion is proposed to tell which trained model is better suited for evaluating the given policy.At last,the effectiveness of the proposed criterion is demonstrated on both benchmark and synthetic offline datasets.展开更多
Objective:To preliminarily construct and apply a longitudinal trajectory model for the prognosis of intracerebral hemorrhage(ICH)based on blood urea nitrogen(BUN)characteristics.Methods:Clinical data from 320 ICH pati...Objective:To preliminarily construct and apply a longitudinal trajectory model for the prognosis of intracerebral hemorrhage(ICH)based on blood urea nitrogen(BUN)characteristics.Methods:Clinical data from 320 ICH patients admitted to our hospital between 2020 and 2024 were collected,including demographic information,National Institutes of Health Stroke Scale(NIHSS)scores at admission,dynamic changes in BUN levels during treatment,and 30-day survival outcomes.A latent class growth model(LCGM)was first used for preliminary modeling,followed by a latent growth mixture modeling(GMM)approach to determine the final model.Three classes of BUN trajectories for ICH prognosis were identified,and latent classes were established.GMM modeling was then performed on these latent classes,considering linear,quadratic,and cubic polynomial forms;six GMM models were constructed and individuals were assigned to latent trajectory groups for validation.Results:LCGM analysis ultimately identified three dynamic BUN trajectory groups:Sustained low-level group(76 cases,23.8%):BUN remained stable between 3.1-9.0 mmol/L,with the highest 30-day survival rate(98.7%).Fluctuating-declining group(222 cases,69.4%):BUN initially increased and then slowly decreased(peak at day 3:15.2 mmol/L),with a 30-day mortality of 8.1%(18/222),higher than the sustained low-level group.Sustained high-level group(22 cases,6.9%):BUN mean>9.0 mmol/L,with a 30-day mortality of 41.7%(P=0.000).GMM model fitting showed that the cubic polynomial GMM model was optimal(AIC=6754.474,BIC=6852.450,Entropy=0.905).Incorporating gender,age,and BMI as covariates revealed significant effects for gender(Estimate=0.045,-0.011,P=0.000,0.000).The AUC for predicting 30-day mortality was 0.88(sensitivity 82.8%,specificity 77.9%),which increased to 0.89 when combined with admission NIHSS scores.Conclusion:The LCGM+GMM model based on dynamic BUN trajectories effectively distinguishes prognostic subgroups in ICH patients.Patients with persistently elevated or fluctuating-rising BUN levels have a significantly higher mortality risk compared to those with sustained low levels.This model provides a new quantitative tool for early identification of high-risk patients and poor prognoses.展开更多
This paper proposes an approach of developing the feature based parametric product modeling system which is suitable for integrated engineering design in CIMS environment.The architecture of ZD--MCADII and the charact...This paper proposes an approach of developing the feature based parametric product modeling system which is suitable for integrated engineering design in CIMS environment.The architecture of ZD--MCADII and the characteristics of its each module are introduced in detail. ZD--MCADII’s product data is managed by an object--oriented database management system OSCAR, and the product model is built according to the standard STEP. The product design is established on a unified product model, and all the product data are globally associated in ZD--MCADII. ZD--MCADII provides various design features to facilitate the product design, and supports the integrity of CAD, CAPP and CAM.展开更多
An improved approach for J-value segmentation （JSEG） is presented for unsupervised color image segmentation. Instead of color quantization algorithm, an automatic classification method based on adaptive mean shift ...An improved approach for J-value segmentation （JSEG） is presented for unsupervised color image segmentation. Instead of color quantization algorithm, an automatic classification method based on adaptive mean shift （AMS） based clustering is used for nonparametric clustering of image data set. The clustering results are used to construct Gaussian mixture modelling （GMM） of image data for the calculation of soft J value. The region growing algorithm used in JSEG is then applied in segmenting the image based on the multiscale soft J-images. Experiments show that the synergism of JSEG and the soft classification based on AMS based clustering and GMM overcomes the limitations of JSEG successfully and is more robust.展开更多
Background:The traditional manual literature screening approach is limited by its time-consuming nature and high labor costs.A pressing issue is how to leverage large language models to enhance the efficiency and qual...Background:The traditional manual literature screening approach is limited by its time-consuming nature and high labor costs.A pressing issue is how to leverage large language models to enhance the efficiency and quality of evidence-based evaluations of drug efficacy and safety.Methods:This study utilized a manually curated reference literature database—comprising vaccine,hypoglycemic agent,and antidepressant evaluation studies—previously developed by our team through conventional systematic review methods.This validated database served as the gold standard for the development and optimization of LitAutoScreener.Following the PICOS(Population,Intervention,Comparison,Outcomes,Study Design)principles,a chain-of-thought reasoning approach with few-shot learning prompts was implemented to develop the screening algorithm.We subsequently evaluated the performance of LitAutoScreener using 2 independent validation cohorts,assessing both classification accuracy and processing efficiency.Results:For respiratory syncytial virus vaccine safety validation title-abstract screening,our tools based on GPT(GPT-4o),Kimi(moonshot-v1-128k),and DeepSeek(deepseek-chat 2.5)demonstrated high accuracy in inclusion/exclusion decisions(99.38%,98.94%,and 98.85%,respectively).Recall rates were 100.00%,99.13%,and 98.26%,with statistically significant performance differences(χ^(2)=5.99,P=0.048),where GPT outperformed the other models.Exclusion reason concordance rates were 98.85%,94.79%,and 96.47%(χ^(2)=30.22,P<0.001).In full-text screening,all models maintained perfect recall(100.00%),with accuracies of 100.00%(GPT),100.00%(Kimi),and 99.45%(DeepSeek).Processing times averaged 1 to 5 s per article for title–abstract screening and 60 s for full-text processing(including PDF preprocessing).Conclusions:LitAutoScreener offers a new approach for efficient literature screening in drug intervention studies,achieving high accuracy and significantly improving screening efficiency.展开更多
Realistic modeling and rendering of dynamic tornado scene is recognized as a challenging task for researchers of computer graphics. In this paper a new physically based method for simulating and animating tornado scen...Realistic modeling and rendering of dynamic tornado scene is recognized as a challenging task for researchers of computer graphics. In this paper a new physically based method for simulating and animating tornado scene is presented. We first propose a Two-Fluid model based on the physical theory of tornado, then we simulate the flow of tornado and its interaction with surrounding objects such as debris, etc. Taking the scattering and absorption of light by the participating media into account, the illumination effects of the tornado scene can be generated realistically. With the support of graphics hardware, various kinds of dynamic tornado scenes can be rendered at interactive rates.展开更多
The enantiomers of chiral drugs often have different potencies, toxicities, and biochemical properties. Therefore, the FDA and other worldwide regulatory agencies require manufactures to test and prove the enantiomeri...The enantiomers of chiral drugs often have different potencies, toxicities, and biochemical properties. Therefore, the FDA and other worldwide regulatory agencies require manufactures to test and prove the enantiomeric purity of chiral drugs. Amino acid based molecular micelles (AABMM) have been used in chiral CE separations since the 1990’s because of their low environmental impact and because their properties can easily be tuned by changing the amino acids in the chiral surfactant head groups. Using molecular dynamics simulations to investigate the structures and properties of AABMM is part of an ongoing study focusing on investigating and elucidating the factors responsible for chiral recognition with AABMM. The results will be useful for the proper design and selection of more efficient chiral selectors. The micelles investigated contained approximately twenty covalently linked surfactant monomers. Each monomer was in turn composed of an undecyl hydrocarbon chain bound to a dipeptide headgroup containing of all combinations of L-Alanine, L-Valine, and L-Leucine. These materials are of interest because they are effective chiral selectors in capillary electrophoresis separations. Molecular dynamics simulation analyses were used to investigate how the sizes and positions of the headgroup amino acid R-groups affected the solvent accessible surface areas of each AABMM chiral center. In addition, headgroup dihedral angle analyses were used to investigate how amino acid R-group size and position affected the overall headgroup conformations. Finally, distance measurements were used to study the structural and conformational flexibilities of each AABMM headgroup. All analyses were performed in the context of a broader study focused on developing structure-based predictive tools to identify the factors responsible for a) self-assembly, b) function, c) higher ordered structure and d) molecular recognition of these amino acid based molecular micelles.展开更多
A more general narrowband regular-shaped geometry-based statistical model(RS-GBSM) combined with the line of sight(LoS) and single bounce(SB) rays for unmanned aerial vehicle(UAV) multiple-input multiple-output(MIMO) ...A more general narrowband regular-shaped geometry-based statistical model(RS-GBSM) combined with the line of sight(LoS) and single bounce(SB) rays for unmanned aerial vehicle(UAV) multiple-input multiple-output(MIMO) channel is proposed in this paper. The channel characteristics, including space-time correlation function(STCF), Doppler power spectral density(DPSD), level crossing rate(LCR) and average fade duration(AFD), are derived based on the single sphere reference model for a non-isotropic environment. The corresponding sum-of-sinusoids(SoS) simulation models including both the deterministic model and statistical model with finite scatterers are also proposed for practicable implementation. The simulation results illustrate that the simulation models well reproduce the channel characteristics of the single sphere reference model with sufficient simulation scatterers. And the statistical model has a better approximation of the reference model in comparison with the deterministic one when the simulation trials of the stochastic model are sufficient. The effects of the parameters such as flight height, moving direction and Rice factor on the characteristics are also studied.展开更多
Metal organic chenlical vapor deposition （AIOCVD） growth systems arc one of the. main types of equipment used for growing single crystal materials, such as GaN. To obtain fihn epitaxial materials with uniform perfor...Metal organic chenlical vapor deposition （AIOCVD） growth systems arc one of the. main types of equipment used for growing single crystal materials, such as GaN. To obtain fihn epitaxial materials with uniform performanee, the flow field and ternperature field in a GaN-MOCVD reactor are investigated by modeling and simulating. To make the simulation results more consistent with the actual situation, the gases in the reactor are considered to be compressible, making it possible to investigate the distributions of gas density and pressure in the reactor. The computational fluid dynamics method is used to stud,v the effects of inlet gas flow velocity, pressure in the reactor, rotational speed of graphite susceptor, and gases used in the growth, which has great guiding~ significance for the growth of GaN fihn materials.展开更多
Surrogate-Based Optimization(SBO) is becoming increasingly popular since it can remarkably reduce the computational cost for design optimizations based on high-fidelity and expensive numerical analyses. However, for c...Surrogate-Based Optimization(SBO) is becoming increasingly popular since it can remarkably reduce the computational cost for design optimizations based on high-fidelity and expensive numerical analyses. However, for complicated optimization problems with a large design space, many design variables, and strong nonlinearity, SBO converges slowly and shows imperfection in local exploitation. This paper proposes a trust region method within the framework of an SBO process based on the Kriging model. In each refinement cycle, new samples are selected by a certain design of experiment method within a variable design space, which is sequentially updated by the trust region method. A multi-dimensional trust-region radius is proposed to improve the adaptability of the developed methodology. Further, the scale factor and the limit factor of the trust region are studied to evaluate their effects on the optimization process. Thereafter, different SBO methods using error-based exploration, prediction-based exploitation, refinement based on the expected improvement function, a hybrid refinement strategy, and the developed trust-regionbased refinement are utilized in four analytical tests. Further, the developed optimization methodology is employed in the drag minimization of an RAE2822 airfoil. Results indicate that it has better robustness and local exploitation capability in comparison with those of other SBO展开更多
A general mathematical model of carrier-based aircraft ski jump take-off is derived based on tensor. The carrier, the aircraft body and the movable parts of the landing gears are treated as independent entities. These...A general mathematical model of carrier-based aircraft ski jump take-off is derived based on tensor. The carrier, the aircraft body and the movable parts of the landing gears are treated as independent entities. These entities are assembled into a multi-rigid-body system with flexible links. Dynamical equations of each entity are derived on the basis of the Newton law and the Euler transformation. Using the invariance property of the tensor, the dynamical and kinematical equations are converted to tensor forms which are invariant under time-dependent coordinate transformations. Then the tensor-formed equations are expressed by the matrix operation. Differential equation group of the matrix form is formulated for the programming. The closure of the model is discussed, and the simulation results are given.展开更多
The design work of motional cable in products is vital due to the difficulty in estimating the potential issues in current researches.In this paper,a physics-based modeling and simulation method for the motional cable...The design work of motional cable in products is vital due to the difficulty in estimating the potential issues in current researches.In this paper,a physics-based modeling and simulation method for the motional cable harness design is presented.The model,based on continuum mechanics,is established by analyzing the force of microelement in equilibrium.During the analysis procedure,three coordinate systems：inertial,Frenet and main-axis coordinate systems are used.By variable substitution and dimensionless processing,the equation set is discretized by differential quadrature method and subsequently becomes an overdetermined nonlinear equation set with boundary conditions solved by Levenberg-Marquardt method.With the profile of motional cable harness obtained from the integral of arithmetic solution,a motion simulation system based on＂path＂and＂profile＂as well as the experimental equipments is built.Using the same parameters as input for the simulation and the real cable harness correspondingly,the issue in designing,such as collision,can be easily found by the simulation system.This research obtains a better result which has no potential collisions by redesign,and the proposed method can be used as an accurate and efficient way in motional cable harness design work.展开更多
In this paper, the structure characteristics of open complex giant systems are concretely analysed in depth, thus the view and its significance to support the meta synthesis engineering with manifold knowledge models...In this paper, the structure characteristics of open complex giant systems are concretely analysed in depth, thus the view and its significance to support the meta synthesis engineering with manifold knowledge models are clarified. Furthermore, the knowledge based multifaceted modeling methodology for open complex giant systems is emphatically studied. The major points are as follows: (1) nonlinear mechanism and general information partition law; (2) from the symmetry and similarity to the acquisition of construction knowledge; (3) structures for hierarchical and nonhierarchical organizations; (4) the integration of manifold knowledge models; (5) the methodology of knowledge based multifaceted modeling.展开更多
文摘Federated Learning(FL)enables joint training over distributed devices without data exchange but is highly vulnerable to attacks by adversaries in the form of model poisoning and malicious update injection.This work proposes Secured-FL,a blockchain-based defensive framework that combines smart contract-based authentication,clustering-driven outlier elimination,and dynamic threshold adjustment to defend against adversarial attacks.The framework was implemented on a private Ethereum network with a Proof-of-Authority consensus algorithm to ensure tamper-resistant and auditable model updates.Large-scale simulation on the Cyber Data dataset,under up to 50%malicious client settings,demonstrates Secured-FL achieves 6%-12%higher accuracy,9%-15%lower latency,and approximately 14%less computational expense compared to the PPSS benchmark framework.Additional tests,including confusion matrices,ROC and Precision-Recall curves,and ablation tests,confirm the interpretability and robustness of the defense.Tests for scalability also show consistent performance up to 500 clients,affirming appropriateness to reasonably large deployments.These results make Secured-FL a feasible,adversarially resilient FL paradigm with promising potential for application in smart cities,medicine,and other mission-critical IoT deployments.
文摘This review explores the use of agent-based modeling(ABM)within the framework of study human emotion and cognition in the context of its ability to simulate complex social interactions,adaptive changes,and evolutionary processes.By representing agents and their defined environments with probabilistic interactions,ABM allows the assessment of the effects of individual behavior at the micro level on the greater social phenomena at the macro level.The review looks into the applications of ABM in portraying some of the key components of emotions and cognition-empathy,cooperation,decision making,and emotional transmission-and analyzes the problems including scalability,empirical validation,and description of sensitive emotional states.The most important conclusion is that merging ABM with information neurobiological data and artificial intelligence(AI)techniques would allow for deepening the interactions within the system and enhancing its responsiveness to stimuli.This review highlights approaches that aim to exploit the ABM methodology more fully and integrates methods from biology,neuroscience,and engineering.This integration could contribute to our understanding of the human behavior evolution and adaptation within systems relevant to policymaking,healthcare,and education.
文摘Conversational recommender systems(CRSs)focus on refining preferences and providing personalized recommendations through natural language interactions and dialogue history.Large language models(LLMs)have shown outstanding performance across various domains,thereby prompting researchers to investigate their applicability in recommendation systems.However,due to the lack of task-specific knowledge and an inefficient feature extraction process,LLMs still have suboptimal performance in recommendation tasks.Therefore,external knowledge sources,such as knowledge graphs(KGs)and knowledge bases(KBs),are often introduced to address the issue of data sparsity.Compared to KGs,KBs possess higher retrieval efficiency,making them more suitable for scenarios where LLMs serve as recommenders.To this end,we introduce a novel framework integrating LLMs with KBs for enhanced retrieval generation,namely LLMKB.LLMKB initially leverages structured knowledge to create mapping dictionaries,extracting entity-relation information from heterogeneous knowledge to construct KBs.Then,LLMKB achieves the embedding calibration between user information representations and documents in KBs through retrieval model fine-tuning.Finally,LLMKB employs retrievalaugmented generation to produce recommendations based on fused text inputs,followed by post-processing.Experiment results on two public CRS datasets demonstrate the effectiveness of our framework.Our code is publicly available at the link:https://anonymous.4open.science/r/LLMKB-6FD0.
文摘Tau plays a crucial role in several neurodegenerative diseases,collectively referred to as tauopathies.Therefore,targeting potential pathological changes in tau could enable useful therapeutic interventions.However,tau is not an easy target because it dynamically interacts with microtubules and other cellular components,which presents a challenge for tau-targeted drugs.New cellular models could aid the development of mechanism-based tau-targeted therapies.
基金supported by the National Natural Science Foundation of China(62473020).
文摘Dear Editor,This letter presents a novel approach to the data-driven control of unknown nonlinear systems.By leveraging online sparse identification based on the Koopman operator,a high-dimensional linear system model approximating the actual system is obtained online.The upper bound of the discrepancy between the identified model and the actual system is estimated using real-time prediction error,which is then utilized in the design of a tube-based robust model predictive controller.The effectiveness of the proposed approach is validated by numerical simulation.
文摘Blend based polymer nanocomposites, comprising Janus nanoparticles at their polymer/polymer interface, were analytically/experimentally evaluated. The modeling procedure was performed in two stages: first, modeling of polymer/polymer interface region comprising Janus nanoparticles and second, modeling of the entire systems as a function of the variation of the blend morphology.In the first stage, the modeling procedure was performed based on the development of the model proposed by Ji et al. and in the second stage, the fundamental of Kolarik's model was used in order to propose a developed and more practical model. It was shown that Janus nanoparticles may form dual polymer/particle interphase at polymer/polymer interface which can drastically affect the final mechanical properties of the system. Comparing the results of tensile tests imposed on different prepared samples with the predictions of the model proved its accuracy and reliability(error < 9%).
文摘Offline policy evaluation,evaluating and selecting complex policies for decision-making by only using offline datasets is important in reinforcement learning.At present,the model-based offline policy evaluation(MBOPE)is widely welcomed because of its easy to implement and good performance.MBOPE directly approximates the unknown value of a given policy using the Monte Carlo method given the estimated transition and reward functions of the environment.Usually,multiple models are trained,and then one of them is selected to be used.However,a challenge remains in selecting an appropriate model from those trained for further use.The authors first analyse the upper bound of the difference between the approximated value and the unknown true value.Theoretical results show that this difference is related to the trajectories generated by the given policy on the learnt model and the prediction error of the transition and reward functions at these generated data points.Based on the theoretical results,a new criterion is proposed to tell which trained model is better suited for evaluating the given policy.At last,the effectiveness of the proposed criterion is demonstrated on both benchmark and synthetic offline datasets.
文摘Objective:To preliminarily construct and apply a longitudinal trajectory model for the prognosis of intracerebral hemorrhage(ICH)based on blood urea nitrogen(BUN)characteristics.Methods:Clinical data from 320 ICH patients admitted to our hospital between 2020 and 2024 were collected,including demographic information,National Institutes of Health Stroke Scale(NIHSS)scores at admission,dynamic changes in BUN levels during treatment,and 30-day survival outcomes.A latent class growth model(LCGM)was first used for preliminary modeling,followed by a latent growth mixture modeling(GMM)approach to determine the final model.Three classes of BUN trajectories for ICH prognosis were identified,and latent classes were established.GMM modeling was then performed on these latent classes,considering linear,quadratic,and cubic polynomial forms;six GMM models were constructed and individuals were assigned to latent trajectory groups for validation.Results:LCGM analysis ultimately identified three dynamic BUN trajectory groups:Sustained low-level group(76 cases,23.8%):BUN remained stable between 3.1-9.0 mmol/L,with the highest 30-day survival rate(98.7%).Fluctuating-declining group(222 cases,69.4%):BUN initially increased and then slowly decreased(peak at day 3:15.2 mmol/L),with a 30-day mortality of 8.1%(18/222),higher than the sustained low-level group.Sustained high-level group(22 cases,6.9%):BUN mean>9.0 mmol/L,with a 30-day mortality of 41.7%(P=0.000).GMM model fitting showed that the cubic polynomial GMM model was optimal(AIC=6754.474,BIC=6852.450,Entropy=0.905).Incorporating gender,age,and BMI as covariates revealed significant effects for gender(Estimate=0.045,-0.011,P=0.000,0.000).The AUC for predicting 30-day mortality was 0.88(sensitivity 82.8%,specificity 77.9%),which increased to 0.89 when combined with admission NIHSS scores.Conclusion:The LCGM+GMM model based on dynamic BUN trajectories effectively distinguishes prognostic subgroups in ICH patients.Patients with persistently elevated or fluctuating-rising BUN levels have a significantly higher mortality risk compared to those with sustained low levels.This model provides a new quantitative tool for early identification of high-risk patients and poor prognoses.
文摘This paper proposes an approach of developing the feature based parametric product modeling system which is suitable for integrated engineering design in CIMS environment.The architecture of ZD--MCADII and the characteristics of its each module are introduced in detail. ZD--MCADII’s product data is managed by an object--oriented database management system OSCAR, and the product model is built according to the standard STEP. The product design is established on a unified product model, and all the product data are globally associated in ZD--MCADII. ZD--MCADII provides various design features to facilitate the product design, and supports the integrity of CAD, CAPP and CAM.
文摘An improved approach for J-value segmentation （JSEG） is presented for unsupervised color image segmentation. Instead of color quantization algorithm, an automatic classification method based on adaptive mean shift （AMS） based clustering is used for nonparametric clustering of image data set. The clustering results are used to construct Gaussian mixture modelling （GMM） of image data for the calculation of soft J value. The region growing algorithm used in JSEG is then applied in segmenting the image based on the multiscale soft J-images. Experiments show that the synergism of JSEG and the soft classification based on AMS based clustering and GMM overcomes the limitations of JSEG successfully and is more robust.
基金supported by the National Natural Science Foundation of China(grant no.72074011)the Research Project of China Drug Supervision and Administration Research Association(2025-Y-Y-012).
文摘Background:The traditional manual literature screening approach is limited by its time-consuming nature and high labor costs.A pressing issue is how to leverage large language models to enhance the efficiency and quality of evidence-based evaluations of drug efficacy and safety.Methods:This study utilized a manually curated reference literature database—comprising vaccine,hypoglycemic agent,and antidepressant evaluation studies—previously developed by our team through conventional systematic review methods.This validated database served as the gold standard for the development and optimization of LitAutoScreener.Following the PICOS(Population,Intervention,Comparison,Outcomes,Study Design)principles,a chain-of-thought reasoning approach with few-shot learning prompts was implemented to develop the screening algorithm.We subsequently evaluated the performance of LitAutoScreener using 2 independent validation cohorts,assessing both classification accuracy and processing efficiency.Results:For respiratory syncytial virus vaccine safety validation title-abstract screening,our tools based on GPT(GPT-4o),Kimi(moonshot-v1-128k),and DeepSeek(deepseek-chat 2.5)demonstrated high accuracy in inclusion/exclusion decisions(99.38%,98.94%,and 98.85%,respectively).Recall rates were 100.00%,99.13%,and 98.26%,with statistically significant performance differences(χ^(2)=5.99,P=0.048),where GPT outperformed the other models.Exclusion reason concordance rates were 98.85%,94.79%,and 96.47%(χ^(2)=30.22,P<0.001).In full-text screening,all models maintained perfect recall(100.00%),with accuracies of 100.00%(GPT),100.00%(Kimi),and 99.45%(DeepSeek).Processing times averaged 1 to 5 s per article for title–abstract screening and 60 s for full-text processing(including PDF preprocessing).Conclusions:LitAutoScreener offers a new approach for efficient literature screening in drug intervention studies,achieving high accuracy and significantly improving screening efficiency.
基金Project supported by the National Basic Research Program (973) of China (No. 2002CB312101) and the National Natural Science Foundation of China (No. 60475013)
文摘Realistic modeling and rendering of dynamic tornado scene is recognized as a challenging task for researchers of computer graphics. In this paper a new physically based method for simulating and animating tornado scene is presented. We first propose a Two-Fluid model based on the physical theory of tornado, then we simulate the flow of tornado and its interaction with surrounding objects such as debris, etc. Taking the scattering and absorption of light by the participating media into account, the illumination effects of the tornado scene can be generated realistically. With the support of graphics hardware, various kinds of dynamic tornado scenes can be rendered at interactive rates.
文摘The enantiomers of chiral drugs often have different potencies, toxicities, and biochemical properties. Therefore, the FDA and other worldwide regulatory agencies require manufactures to test and prove the enantiomeric purity of chiral drugs. Amino acid based molecular micelles (AABMM) have been used in chiral CE separations since the 1990’s because of their low environmental impact and because their properties can easily be tuned by changing the amino acids in the chiral surfactant head groups. Using molecular dynamics simulations to investigate the structures and properties of AABMM is part of an ongoing study focusing on investigating and elucidating the factors responsible for chiral recognition with AABMM. The results will be useful for the proper design and selection of more efficient chiral selectors. The micelles investigated contained approximately twenty covalently linked surfactant monomers. Each monomer was in turn composed of an undecyl hydrocarbon chain bound to a dipeptide headgroup containing of all combinations of L-Alanine, L-Valine, and L-Leucine. These materials are of interest because they are effective chiral selectors in capillary electrophoresis separations. Molecular dynamics simulation analyses were used to investigate how the sizes and positions of the headgroup amino acid R-groups affected the solvent accessible surface areas of each AABMM chiral center. In addition, headgroup dihedral angle analyses were used to investigate how amino acid R-group size and position affected the overall headgroup conformations. Finally, distance measurements were used to study the structural and conformational flexibilities of each AABMM headgroup. All analyses were performed in the context of a broader study focused on developing structure-based predictive tools to identify the factors responsible for a) self-assembly, b) function, c) higher ordered structure and d) molecular recognition of these amino acid based molecular micelles.
基金supported in part by the National Natural Science Foundation of China under Grant 61622101 and Grant 61571020National Science and Technology Major Project under Grant 2018ZX03001031
文摘A more general narrowband regular-shaped geometry-based statistical model(RS-GBSM) combined with the line of sight(LoS) and single bounce(SB) rays for unmanned aerial vehicle(UAV) multiple-input multiple-output(MIMO) channel is proposed in this paper. The channel characteristics, including space-time correlation function(STCF), Doppler power spectral density(DPSD), level crossing rate(LCR) and average fade duration(AFD), are derived based on the single sphere reference model for a non-isotropic environment. The corresponding sum-of-sinusoids(SoS) simulation models including both the deterministic model and statistical model with finite scatterers are also proposed for practicable implementation. The simulation results illustrate that the simulation models well reproduce the channel characteristics of the single sphere reference model with sufficient simulation scatterers. And the statistical model has a better approximation of the reference model in comparison with the deterministic one when the simulation trials of the stochastic model are sufficient. The effects of the parameters such as flight height, moving direction and Rice factor on the characteristics are also studied.
基金Supported by the National Key R&D Program of China under Grant No 2016YFB0400104
文摘Metal organic chenlical vapor deposition （AIOCVD） growth systems arc one of the. main types of equipment used for growing single crystal materials, such as GaN. To obtain fihn epitaxial materials with uniform performanee, the flow field and ternperature field in a GaN-MOCVD reactor are investigated by modeling and simulating. To make the simulation results more consistent with the actual situation, the gases in the reactor are considered to be compressible, making it possible to investigate the distributions of gas density and pressure in the reactor. The computational fluid dynamics method is used to stud,v the effects of inlet gas flow velocity, pressure in the reactor, rotational speed of graphite susceptor, and gases used in the growth, which has great guiding~ significance for the growth of GaN fihn materials.
基金co-supported by the National Natural Science Foundation of China (No. 11502209)the Free Research Projects of the Central University Funding of China (No. 3102015ZY007)
文摘Surrogate-Based Optimization(SBO) is becoming increasingly popular since it can remarkably reduce the computational cost for design optimizations based on high-fidelity and expensive numerical analyses. However, for complicated optimization problems with a large design space, many design variables, and strong nonlinearity, SBO converges slowly and shows imperfection in local exploitation. This paper proposes a trust region method within the framework of an SBO process based on the Kriging model. In each refinement cycle, new samples are selected by a certain design of experiment method within a variable design space, which is sequentially updated by the trust region method. A multi-dimensional trust-region radius is proposed to improve the adaptability of the developed methodology. Further, the scale factor and the limit factor of the trust region are studied to evaluate their effects on the optimization process. Thereafter, different SBO methods using error-based exploration, prediction-based exploitation, refinement based on the expected improvement function, a hybrid refinement strategy, and the developed trust-regionbased refinement are utilized in four analytical tests. Further, the developed optimization methodology is employed in the drag minimization of an RAE2822 airfoil. Results indicate that it has better robustness and local exploitation capability in comparison with those of other SBO
文摘A general mathematical model of carrier-based aircraft ski jump take-off is derived based on tensor. The carrier, the aircraft body and the movable parts of the landing gears are treated as independent entities. These entities are assembled into a multi-rigid-body system with flexible links. Dynamical equations of each entity are derived on the basis of the Newton law and the Euler transformation. Using the invariance property of the tensor, the dynamical and kinematical equations are converted to tensor forms which are invariant under time-dependent coordinate transformations. Then the tensor-formed equations are expressed by the matrix operation. Differential equation group of the matrix form is formulated for the programming. The closure of the model is discussed, and the simulation results are given.
基金Supported by National Natural Science Foundation of China(Grant No.51275047)
文摘The design work of motional cable in products is vital due to the difficulty in estimating the potential issues in current researches.In this paper,a physics-based modeling and simulation method for the motional cable harness design is presented.The model,based on continuum mechanics,is established by analyzing the force of microelement in equilibrium.During the analysis procedure,three coordinate systems：inertial,Frenet and main-axis coordinate systems are used.By variable substitution and dimensionless processing,the equation set is discretized by differential quadrature method and subsequently becomes an overdetermined nonlinear equation set with boundary conditions solved by Levenberg-Marquardt method.With the profile of motional cable harness obtained from the integral of arithmetic solution,a motion simulation system based on＂path＂and＂profile＂as well as the experimental equipments is built.Using the same parameters as input for the simulation and the real cable harness correspondingly,the issue in designing,such as collision,can be easily found by the simulation system.This research obtains a better result which has no potential collisions by redesign,and the proposed method can be used as an accurate and efficient way in motional cable harness design work.
文摘In this paper, the structure characteristics of open complex giant systems are concretely analysed in depth, thus the view and its significance to support the meta synthesis engineering with manifold knowledge models are clarified. Furthermore, the knowledge based multifaceted modeling methodology for open complex giant systems is emphatically studied. The major points are as follows: (1) nonlinear mechanism and general information partition law; (2) from the symmetry and similarity to the acquisition of construction knowledge; (3) structures for hierarchical and nonhierarchical organizations; (4) the integration of manifold knowledge models; (5) the methodology of knowledge based multifaceted modeling.
