The production mode of manufacturing industry presents characteristics of multiple varieties,small-batch and personalization,leading to frequent disturbances in workshop.Traditional centralized scheduling methods are ...The production mode of manufacturing industry presents characteristics of multiple varieties,small-batch and personalization,leading to frequent disturbances in workshop.Traditional centralized scheduling methods are difficult to achieve efficient and real-time production management under dynamic disturbance.In order to improve the intelligence and adaptability of production scheduler,a novel distributed scheduling architecture is proposed,which has the ability to autonomously allocate tasks and handle disturbances.All production tasks are scheduled through autonomous collaboration and decision-making between intelligent machines.Firstly,the multi-agent technology is applied to build a self-organizing manufacturing system,enabling each machine to be equipped with the ability of active information interaction and joint-action execution.Secondly,various self-organizing collaboration strategies are designed to effectively facilitate cooperation and competition among multiple agents,thereby flexibly achieving global perception of environmental state.To ensure the adaptability and superiority of production decisions in dynamic environment,deep reinforcement learning is applied to build a smart production scheduler:Based on the perceived environment state,the scheduler intelligently generates the optimal production strategy to guide the task allocation and resource configuration.The feasibility and effectiveness of the proposed method are verified through three experimental scenarios using a discrete manufacturing workshop as the test bed.Compared to heuristic dispatching rules,the proposed method achieves an average performance improvement of 34.0%in three scenarios in terms of order tardiness.The proposed system can provide a new reference for the design of smart manufacturing systems.展开更多
The modern war features a highly distributed coordination. In the face of great time constrains, it is important to change command organizations to adapt to the real environment. Therefore it's a key step to set u...The modern war features a highly distributed coordination. In the face of great time constrains, it is important to change command organizations to adapt to the real environment. Therefore it's a key step to set up adaptive C2 teams. In this paper, the relational problems about distributed C2 organizational structure adaptation are discussed, and the methodology for team decision making design based on the object oriented technique is studied.展开更多
In this paper,an adaptive cubic regularisation algorithm based on affine scaling methods(ARCBASM)is proposed for solving nonlinear equality constrained programming with nonnegative constraints on variables.From the op...In this paper,an adaptive cubic regularisation algorithm based on affine scaling methods(ARCBASM)is proposed for solving nonlinear equality constrained programming with nonnegative constraints on variables.From the optimality conditions of the problem,we introduce appropriate affine matrix and construct an affine scaling ARC subproblem with linearized constraints.Composite step methods and reduced Hessian methods are applied to tackle the linearized constraints.As a result,a standard unconstrained ARC subproblem is deduced and its solution can supply sufficient decrease.The fraction to the boundary rule maintains the strict feasibility(for nonnegative constraints on variables)of every iteration point.Reflection techniques are employed to prevent the iterations from approaching zero too early.Under mild assumptions,global convergence of the algorithm is analysed.Preliminary numerical results are reported.展开更多
Background:Despite the promise shown by large language models(LLMs)for standardized tasks,their multidimensional performance in real-world oncology decision-making remains unevaluated.This study aims to introduce a fr...Background:Despite the promise shown by large language models(LLMs)for standardized tasks,their multidimensional performance in real-world oncology decision-making remains unevaluated.This study aims to introduce a framework for evaluating LLMs and physician decisions in challenging lung cancer cases.Methods:We curated 50 challenging lung cancer cases(25 local and 25 published)classified as complex,rare,or refractory.Blinded three-dimensional,five-point Likert evaluations(1–5 for comprehensiveness,specificity,and readability)compared standalone LLMs(DeepSeek R1,Claude 3.5,Gemini 1.5,and GPT-4o),physicians by experience level(junior,intermediate,and senior),and AI-assisted juniors;intergroup differences and augmentation effects were analyzed statistically.Results:Of 50 challenging cases(18 complex,17 rare,and 15 refractory)rated by three experts,DeepSeek R1 achieved scores of 3.95±0.33,3.71±0.53,and 4.26±0.18 for comprehensiveness,specificity,and readability,respectively,positioning it between intermediate(3.68,3.68,3.75)and senior(4.50,4.64,4.53)physicians.GPT-4o and Claude 3.5 reached intermediate physician–level comprehensiveness(3.76±0.39,3.60±0.39)but junior-to-intermediate physician–level specificity(3.39±0.39,3.39±0.49).All LLMs scored higher on rare cases than intermediate physicians but fell below junior physicians in refractory-case specificity.AIassisted junior physicians showed marked gains in rare cases,with comprehensiveness rising from 2.32 to 4.29(84.8%),specificity from 2.24 to 4.26(90.8%),and readability from 2.76 to 4.59(66.0%),while specificity declined by 3.2%(3.17 to 3.07)in refractory cases.Error analysis showed complementary strengths,with physicians demonstrating reasoning stability and LLMs excelling in knowledge updating and risk management.Conclusions:LLMs performed variably in clinical decision-making tasks depending on case type,performing better in rare cases and worse in refractory cases requiring longitudinal reasoning.Complementary strengths between LLMs and physicians support case-and task-tailored human–AI collaboration.展开更多
Environmental problems are intensifying due to the rapid growth of the population,industry,and urban infrastructure.This expansion has resulted in increased air and water pollution,intensified urban heat island effect...Environmental problems are intensifying due to the rapid growth of the population,industry,and urban infrastructure.This expansion has resulted in increased air and water pollution,intensified urban heat island effects,and greater runoff from parks and other green spaces.Addressing these challenges requires prioritizing green infrastructure and other sustainable urban development strategies.This study introduces a novel Integrated Decision Support System that combines Pythagorean Fuzzy Sets with the Advanced Alternative Ranking Order Method allowing for Two-Step Normalization(AAROM-TN),enhanced by a dual weighting strategy.The weighting approach integrates the Criteria Importance Through Intercriteria Correlation(CRITIC)method with the Criteria Importance through Means and Standard Deviation(CIMAS)technique.The originality of the proposed framework lies in its ability to objectively quantify criteria importance using CRITIC,incorporate decision-makers’preferences through CIMAS,and capture the uncertainty and hesitation inherent in human judgment via Pythagorean Fuzzy Sets.A case study evaluating green infrastructure alternatives in metropolitan regions demonstrates the applicability and effectiveness of the framework.A sensitivity analysis is conducted to examine how variations in criteria weights affect the rankings and to evaluate the robustness of the results.Furthermore,a comparative analysis highlights the practical and financial implications of each alternative by assessing their respective strengths and weaknesses.展开更多
Dear Editor,This letter deals with the autonomous underwater vehicle(AUV)three dimensional(3D)trajectory tracking control chronically suffering from poor accuracy and efficiency in complex hydrodynamics.A state-of-the...Dear Editor,This letter deals with the autonomous underwater vehicle(AUV)three dimensional(3D)trajectory tracking control chronically suffering from poor accuracy and efficiency in complex hydrodynamics.A state-of-the-art predictive adaptive controller(PAC)is proposed with a distinct dual closed-loop structure.展开更多
Starting from the foundational static traits underlying the growth and development of flue-cured tobacco, this research conducts a systematic examination of the phenomena and theoretical principles associated with env...Starting from the foundational static traits underlying the growth and development of flue-cured tobacco, this research conducts a systematic examination of the phenomena and theoretical principles associated with environment-driven adaptive changes during its cultivation. It was found that environmental variables-including temperature, light, and moisture-elicit directional shifts in static traits ( e.g. , chemical composition, morphological architecture, and leaf tissue structure) toward enhanced environmental adaptation, characterized by graduality, juvenility, similarity, and correlativity. Upon alterations in ambient conditions, flue-cured tobacco modulates its static traits through integrated physical, chemical, and biological-genetic mechanisms, aiming to optimize resource utilization, mitigate environmental constraints, and preserve internal homeostasis alongside metabolic balance. The investigation further reveals that the adaptive scope of flue-cured tobacco to field environments is malleable and can be extended and elevated via adaptive conditioning commencing at the juvenile stage. In addition, the adaptive alignment between static traits and environmental parameters exerts a substantial impact on the plant s growth dynamics, yield performance, and quality attributes. Beyond its relevance to flue-cured tobacco, the proposed theory offers a meaningful framework for elucidating the pervasive adaptive strategies employed by plants and broader biological systems in response to environmental contingencies.展开更多
With the rapid development of artificial intelligence,intelligent air combat maneuver decision-making(ACMD)has garnered global attention.Although deep reinforcement learning provides a promising approach to ACMD,exist...With the rapid development of artificial intelligence,intelligent air combat maneuver decision-making(ACMD)has garnered global attention.Although deep reinforcement learning provides a promising approach to ACMD,existing methods often suffer from rigid reward functions and limited adaptability to evolving adversarial strategies.Moreover,most research assumes open airspace,overlooking the influence of potential obstacles.In this paper,we address one-on-one within-visual-range ACMD in obstructed environments,and propose an improved Soft Actor-Critic(SAC)algorithm trained under a curriculum self-play framework.A maneuver strategy mirroring inference module is integrated to estimate each other's likely positions when visual obstruction occurs.By leveraging curriculum learning to guide progressive experience accumulation and self-play for adversarial evolution,our method enhances both training efficiency and tactical diversity.We further integrate an attention mechanism that dynamically adjusts the weights of sub-rewards,enabling the learned policy to adapt to rapidly changing air combat situations.Numerical simulations demonstrate that our enhanced SAC converges more quickly and achieves higher win rates than other baseline methods.An animation is available at bilibili.com/video/BV1BHVszHE98 for better illustration.展开更多
In federated learning,backdoor attacks have become an important research topic with their wide application in processing sensitive datasets.Since federated learning detects or modifies local models through defense mec...In federated learning,backdoor attacks have become an important research topic with their wide application in processing sensitive datasets.Since federated learning detects or modifies local models through defense mechanisms during aggregation,it is difficult to conduct effective backdoor attacks.In addition,existing backdoor attack methods are faced with challenges,such as low backdoor accuracy,poor ability to evade anomaly detection,and unstable model training.To address these challenges,a method called adaptive simulation backdoor attack(ASBA)is proposed.Specifically,ASBA improves the stability of model training by manipulating the local training process and using an adaptive mechanism,the ability of the malicious model to evade anomaly detection by combing large simulation training and clipping,and the backdoor accuracy by introducing a stimulus model to amplify the impact of the backdoor in the global model.Extensive comparative experiments under five advanced defense scenarios show that ASBA can effectively evade anomaly detection and achieve high backdoor accuracy in the global model.Furthermore,it exhibits excellent stability and effectiveness after multiple rounds of attacks,outperforming state-of-the-art backdoor attack methods.展开更多
With the increasing complexity of industrial automation,planetary gearboxes play a vital role in largescale equipment transmission systems,directly impacting operational efficiency and safety.Traditional maintenance s...With the increasing complexity of industrial automation,planetary gearboxes play a vital role in largescale equipment transmission systems,directly impacting operational efficiency and safety.Traditional maintenance strategies often struggle to accurately predict the degradation process of equipment,leading to excessive maintenance costs or potential failure risks.However,existing prediction methods based on statistical models are difficult to adapt to nonlinear degradation processes.To address these challenges,this study proposes a novel condition-based maintenance framework for planetary gearboxes.A comprehensive full-lifecycle degradation experiment was conducted to collect raw vibration signals,which were then processed using a temporal convolutional network autoencoder with multi-scale perception capability to extract deep temporal degradation features,enabling the collaborative extraction of longperiod meshing frequencies and short-term impact features from the vibration signals.Kernel principal component analysis was employed to fuse and normalize these features,enhancing the characterization of degradation progression.A nonlinear Wiener process was used to model the degradation trajectory,with a threshold decay function introduced to dynamically adjust maintenance strategies,and model parameters optimized through maximum likelihood estimation.Meanwhile,the maintenance strategy was optimized to minimize costs per unit time,determining the optimal maintenance timing and preventive maintenance threshold.The comprehensive indicator of degradation trends extracted by this method reaches 0.756,which is 41.2%higher than that of traditional time-domain features;the dynamic threshold strategy reduces the maintenance cost per unit time to 55.56,which is 8.9%better than that of the static threshold optimization.Experimental results demonstrate significant reductions in maintenance costs while enhancing system reliability and safety.This study realizes the organic integration of deep learning and reliability theory in the maintenance of planetary gearboxes,provides an interpretable solution for the predictive maintenance of complex mechanical systems,and promotes the development of condition-based maintenance strategies for planetary gearboxes.展开更多
Small RNAs(sRNAs)are important non-coding RNAs that usually play crucial roles in gene expression at the post-transcriptional level.The sRNAs have mostly been investigated in model microorganisms such as Escherichia c...Small RNAs(sRNAs)are important non-coding RNAs that usually play crucial roles in gene expression at the post-transcriptional level.The sRNAs have mostly been investigated in model microorganisms such as Escherichia coli and some pathogens.Nevertheless,microbial sRNAs from extreme environments such as the polar regions and deep sea have recently been discovered and analyzed for their unique roles in stress response,metabolic regulation and adaptation to extreme environments.These sRNAs fine-tune gene expression during oxidative and radiation stress,and modulate temperature and osmotic pressure responses.Representative sRNAs and their functions in thermophilic,psychrophilic,halophilic and radiation-tolerant bacteria are summarized in this review.Despite challenges in sample collection,RNA isolation,and functional annotation,the study of sRNAs in extreme environments provides opportunities for discovering novel regulatory mechanisms,applying them to biotechnology,and advancing our understanding of evolutionary adaptations.Looking ahead,high-throughput sequencing,synthetic biology,and multi-omics integration will bring new breakthroughs in discovering novel sRNAs and their functions and regulatory mechanisms.Such advancements are poised to enable comprehensive characterization of sRNA-mediated regulatory networks in extremophiles and unlock their biotechnological potential through mechanism-driven applications.展开更多
Epoxy resins are widely employed in wind turbine blades,drone rotors,and automotive interiors due to their excel-lent mechani-cal proper-ties and long service life.However,their insoluble and infusible cross-linked ne...Epoxy resins are widely employed in wind turbine blades,drone rotors,and automotive interiors due to their excel-lent mechani-cal proper-ties and long service life.However,their insoluble and infusible cross-linked networks pose a significant re-cycling challenge,particularly with the impending retirement of the first generation of wind turbine blades.In this work,we reported a fully bio-based epoxy Vitrimer(FEP)incorporat-ing a dual-dynamic covalent network design and systematically investigated the influence of the 1,5,7-triazabicyclo[4.4.0]dec-5-ene(TBD)catalyst on its curing kinetics,thermal/mechan-ical properties,dynamic exchange behavior,and degradation performance in a mild alkaline solution.Compared to conventional epoxy resins,FEP exhibited superior tensile strength and elongation at break at an optimal TBD concentration(2 wt%),achieving an excellent strength-toughness balance.The presence of TBD accelerated the exchange rates of both disulfide and ester bonds,endowing FEP with notable stress relaxation at elevated tempera-tures.Moreover,FEP demonstrated complete dissolution in 1 mol/L NaOH within 6 h at 25℃.These results underscored the exceptional strength,toughness,and recyclability of FEP,positioning it as a promising,environmentally friendly matrix resin for next-generation appli-cations in the new energy sector.展开更多
In this paper,we propose a fuzzy logic-based coded event-triggered control with self-adjustable prescribed performance(FL-CEC-SPP)to address the trade-off between control performance and communication efficiency in re...In this paper,we propose a fuzzy logic-based coded event-triggered control with self-adjustable prescribed performance(FL-CEC-SPP)to address the trade-off between control performance and communication efficiency in resource-constrained networked control systems.The method integrates a fuzzy-coded event-triggered controller into a coded control framework to dynamically adjust the triggering threshold,thereby reducing unnecessary transmissions while maintaining system stability.A self-adjustable prescribed performance constraint is also incorporated to ensure that the tracking error remains within predefined bounds under arbitrary initial conditions.Theoretical analyses and simulation comparisons show that the method proposed in this paper maintains good tracking performance and stability while reducing the communication burden,and has wide applications in resource-constrained network control systems.展开更多
Gait recognition is a key biometric for long-distance identification,yet its performance is severely degraded by real-world challenges such as varying clothing,carrying conditions,and changing viewpoints.While combini...Gait recognition is a key biometric for long-distance identification,yet its performance is severely degraded by real-world challenges such as varying clothing,carrying conditions,and changing viewpoints.While combining silhouette and skeleton data is a promising direction,effectively fusing these heterogeneous modalities and adaptively weighting their contributions in response to diverse conditions remains a central problem.This paper introduces GaitMAFF,a novelMulti-modal Adaptive Feature Fusion Network,to address this challenge.Our approach first transforms discrete skeleton joints into a dense SkeletonMap representation to align with silhouettes,then employs an attention-based module to dynamically learn the fusion weights between the two modalities.These fused features are processed by a powerful spatio-temporal backbone withWeighted Global-Local Feature FusionModules(WFFM)to learn a discriminative representation.Extensive experiments on the challenging CCPG and Gait3D datasets show that GaitMAFF achieves state-of-the-art performance,with an average Rank-1 accuracy of 84.6%on CCPG and 58.7%on Gait3D.These results demonstrate that our adaptive fusion strategy effectively integrates complementary multimodal information,significantly enhancing gait recognition robustness and accuracy in complex scenes and providing a practical solution for real-world applications.展开更多
Designing appropriate loss functions is critical to the success of supervised learning models.However,most conventional losses are fixed and manually designed,making them suboptimal for diverse and dynamic learning sc...Designing appropriate loss functions is critical to the success of supervised learning models.However,most conventional losses are fixed and manually designed,making them suboptimal for diverse and dynamic learning scenarios.In this work,we propose an Adaptive Meta-Loss Network(Adaptive-MLN)that learns to generate taskagnostic loss functions tailored to evolving classification problems.Unlike traditional methods that rely on static objectives,Adaptive-MLN treats the loss function itself as a trainable component,parameterized by a shallow neural network.To enable flexible,gradient-free optimization,we introduce a hybrid evolutionary approach that combines GeneticAlgorithms(GA)for global exploration and Evolution Strategies(ES)for local refinement.This co-evolutionary process dynamically adjusts the loss landscape,improvingmodel generalization without relying on analytic gradients or handcrafted heuristics.Experimental evaluations on synthetic tasks and the CIFAR-10 andMNIST datasets demonstrate that our approach consistently outperforms standard losses such as Cross-Entropy and Mean Squared Error in terms of accuracy,convergence,and adaptability.展开更多
Modern battlefields exhibit high dynamism,where traditional static weighting methods in combat effectiveness assessment fail to capture real-time changes in indicator values,leading to limited assessment accuracy—esp...Modern battlefields exhibit high dynamism,where traditional static weighting methods in combat effectiveness assessment fail to capture real-time changes in indicator values,leading to limited assessment accuracy—especially critical in scenarios like sudden electronic warfare or degraded command,where static weights cannot reflect the operational value decay or surge of key indicators.To address this issue,this study proposes a dynamic adaptive weightingmethod for evaluation indicators based onG1-CRITIC-PIVW.First,theG1(Sequential Relationship Analysis Method)subjective weighting method—translates expert knowledge into indicator importance rankings—leverages expert knowledge to quantify the relative importance of indicators via sequential relationship ranking,while the CRITIC(Criteria Importance Through Intercriteria Correlation)objective weighting method—derives weights from data characteristics by integrating variability and inter-correlations—calculates weights by integrating indicator variability and inter-indicator correlations,ensuring data-driven objectivity.These two sets of weights are then fused using a deviation coefficient optimization model,minimizing the squared deviation from a reference weight and adjusting the fusion coefficient via Spearman’s rank correlation to resolve potential conflicts between subjective and objective judgments.Subsequently,the PIVW(Punishment-Incentive VariableWeight)theory—adapts weights to realtime indicator performance via penalty/incentive rules—is applied for dynamic adjustment.Scenario-specific penalty λ_(1) and incentive λ_(2) thresholds are set based on operational priorities and indicator volatility,penalizing indicators with values below λ_(1) and incentivizing those exceeding λ_(2) to reflect real-time indicator performance.Experimental validation was conducted using an Air Defense and Anti-Missile(ADAM)system effectiveness assessment framework,with data covering 7 indicators across 3 combat scenarios.Results show that compared to static weighting methods,the proposed method reduces MAE(Mean Absolute Error)by 15%-20% and weighted decision error rate by 84.2%,effectively reducing overestimation/underestimation of combat effectiveness in dynamic scenarios;compared to Entropy-TOPSIS,it lowers MAE by 12% while achieving a weighted Kendall’sτconsistency coefficient of 0.85,ensuring higher alignment with expert judgment.This method enhances the accuracy and scenario adaptability of effectiveness assessment,providing reliable decision support for dynamic battlefield environments.展开更多
The rapidly evolving cybersecurity threat landscape exposes a critical flaw in traditional educational programs where static curricula cannot adapt swiftly to novel attack vectors.This creates a significant gap betwee...The rapidly evolving cybersecurity threat landscape exposes a critical flaw in traditional educational programs where static curricula cannot adapt swiftly to novel attack vectors.This creates a significant gap between theoretical knowledge and the practical defensive capabilities needed in the field.To address this,we propose TeachSecure-CTI,a novel framework for adaptive cybersecurity curriculumgeneration that integrates real-time Cyber Threat Intelligence(CTI)with AI-driven personalization.Our framework employs a layered architecture featuring a CTI ingestion and clusteringmodule,natural language processing for semantic concept extraction,and a reinforcement learning agent for adaptive content sequencing.Bydynamically aligning learningmaterialswithboththe evolving threat environment and individual learner profiles,TeachSecure-CTI ensures content remains current,relevant,and tailored.A 12-week study with 150 students across three institutions demonstrated that the framework improves learning gains by 34%,significantly exceeding the 12%–21%reported in recent literature.The system achieved 84.8%personalization accuracy,85.9%recognition accuracy for MITRE ATT&CK tactics,and a 31%faster competency development rate compared to static curricula.These findings have implications beyond academia,extending to workforce development,cyber range training,and certification programs.By bridging the gap between dynamic threats and static educational materials,TeachSecure-CTI offers an empirically validated,scalable solution for cultivating cybersecurity professionals capable of responding to modern threats.展开更多
While reinforcement learning-based underwater acoustic adaptive modulation shows promise for enabling environment-adaptive communication as supported by extensive simulation-based research,its practical performance re...While reinforcement learning-based underwater acoustic adaptive modulation shows promise for enabling environment-adaptive communication as supported by extensive simulation-based research,its practical performance remains underexplored in field investigations.To evaluate the practical applicability of this emerging technique in adverse shallow sea channels,a field experiment was conducted using three communication modes:orthogonal frequency division multiplexing(OFDM),M-ary frequency-shift keying(MFSK),and direct sequence spread spectrum(DSSS)for reinforcement learning-driven adaptive modulation.Specifically,a Q-learning method is used to select the optimal modulation mode according to the channel quality quantified by signal-to-noise ratio,multipath spread length,and Doppler frequency offset.Experimental results demonstrate that the reinforcement learning-based adaptive modulation scheme outperformed fixed threshold detection in terms of total throughput and average bit error rate,surpassing conventional adaptive modulation strategies.展开更多
This paper explores the adaptive exponentially designated-time stabilization issue via event-triggered feedback for a kind of uncertain high-order nonlinear systems.The motivation mainly comes from the following two c...This paper explores the adaptive exponentially designated-time stabilization issue via event-triggered feedback for a kind of uncertain high-order nonlinear systems.The motivation mainly comes from the following two challenges:the undesired singularity problem arising from infinite control gains at the prescribed-time instant,the effective trade-off between the control amplitude and the triggering duration.The goal is to build an event-triggered mechanism comprising a skillful triggered rule alongside a time-dependent threshold.Utilizing the designed control strategy,the solutions'existence and the prevention of Zeno phenomenon are successfully guaranteed by using a new transformation equipped with a time-varying function and redesigning the continuous state-feedback dominance approach with an array of integral functions involving embedded sign functions.Better than existing prescribed-time methods,our approach not only ensures that state variables converge to a small compact set before a designated time and stay there henceforth,and converge to the origin exponentially,but also ensures that the controller continuously works on the whole-time horizon.Two illustrative examples are given to show the effectiveness of the devised scheme.展开更多
In image analysis,high-precision semantic segmentation predominantly relies on supervised learning.Despite significant advancements driven by deep learning techniques,challenges such as class imbalance and dynamic per...In image analysis,high-precision semantic segmentation predominantly relies on supervised learning.Despite significant advancements driven by deep learning techniques,challenges such as class imbalance and dynamic performance evaluation persist.Traditional weighting methods,often based on pre-statistical class counting,tend to overemphasize certain classes while neglecting others,particularly rare sample categories.Approaches like focal loss and other rare-sample segmentation techniques introduce multiple hyperparameters that require manual tuning,leading to increased experimental costs due to their instability.This paper proposes a novel CAWASeg framework to address these limitations.Our approach leverages Grad-CAM technology to generate class activation maps,identifying key feature regions that the model focuses on during decision-making.We introduce a Comprehensive Segmentation Performance Score(CSPS)to dynamically evaluate model performance by converting these activation maps into pseudo mask and comparing them with Ground Truth.Additionally,we design two adaptive weights for each class:a Basic Weight(BW)and a Ratio Weight(RW),which the model adjusts during training based on real-time feedback.Extensive experiments on the COCO-Stuff,CityScapes,and ADE20k datasets demonstrate that our CAWASeg framework significantly improves segmentation performance for rare sample categories while enhancing overall segmentation accuracy.The proposed method offers a robust and efficient solution for addressing class imbalance in semantic segmentation tasks.展开更多
基金supported by the Scientific Research Foundation of Nanjing Institute of Technology(No.YKJ202425)the National Natural Science Foundation of China(No.72301130).
文摘The production mode of manufacturing industry presents characteristics of multiple varieties,small-batch and personalization,leading to frequent disturbances in workshop.Traditional centralized scheduling methods are difficult to achieve efficient and real-time production management under dynamic disturbance.In order to improve the intelligence and adaptability of production scheduler,a novel distributed scheduling architecture is proposed,which has the ability to autonomously allocate tasks and handle disturbances.All production tasks are scheduled through autonomous collaboration and decision-making between intelligent machines.Firstly,the multi-agent technology is applied to build a self-organizing manufacturing system,enabling each machine to be equipped with the ability of active information interaction and joint-action execution.Secondly,various self-organizing collaboration strategies are designed to effectively facilitate cooperation and competition among multiple agents,thereby flexibly achieving global perception of environmental state.To ensure the adaptability and superiority of production decisions in dynamic environment,deep reinforcement learning is applied to build a smart production scheduler:Based on the perceived environment state,the scheduler intelligently generates the optimal production strategy to guide the task allocation and resource configuration.The feasibility and effectiveness of the proposed method are verified through three experimental scenarios using a discrete manufacturing workshop as the test bed.Compared to heuristic dispatching rules,the proposed method achieves an average performance improvement of 34.0%in three scenarios in terms of order tardiness.The proposed system can provide a new reference for the design of smart manufacturing systems.
文摘The modern war features a highly distributed coordination. In the face of great time constrains, it is important to change command organizations to adapt to the real environment. Therefore it's a key step to set up adaptive C2 teams. In this paper, the relational problems about distributed C2 organizational structure adaptation are discussed, and the methodology for team decision making design based on the object oriented technique is studied.
基金Supported by the National Natural Science Foundation of China(12071133)Natural Science Foundation of Henan Province(252300421993)Key Scientific Research Project of Higher Education Institutions in Henan Province(25B110005)。
文摘In this paper,an adaptive cubic regularisation algorithm based on affine scaling methods(ARCBASM)is proposed for solving nonlinear equality constrained programming with nonnegative constraints on variables.From the optimality conditions of the problem,we introduce appropriate affine matrix and construct an affine scaling ARC subproblem with linearized constraints.Composite step methods and reduced Hessian methods are applied to tackle the linearized constraints.As a result,a standard unconstrained ARC subproblem is deduced and its solution can supply sufficient decrease.The fraction to the boundary rule maintains the strict feasibility(for nonnegative constraints on variables)of every iteration point.Reflection techniques are employed to prevent the iterations from approaching zero too early.Under mild assumptions,global convergence of the algorithm is analysed.Preliminary numerical results are reported.
文摘Background:Despite the promise shown by large language models(LLMs)for standardized tasks,their multidimensional performance in real-world oncology decision-making remains unevaluated.This study aims to introduce a framework for evaluating LLMs and physician decisions in challenging lung cancer cases.Methods:We curated 50 challenging lung cancer cases(25 local and 25 published)classified as complex,rare,or refractory.Blinded three-dimensional,five-point Likert evaluations(1–5 for comprehensiveness,specificity,and readability)compared standalone LLMs(DeepSeek R1,Claude 3.5,Gemini 1.5,and GPT-4o),physicians by experience level(junior,intermediate,and senior),and AI-assisted juniors;intergroup differences and augmentation effects were analyzed statistically.Results:Of 50 challenging cases(18 complex,17 rare,and 15 refractory)rated by three experts,DeepSeek R1 achieved scores of 3.95±0.33,3.71±0.53,and 4.26±0.18 for comprehensiveness,specificity,and readability,respectively,positioning it between intermediate(3.68,3.68,3.75)and senior(4.50,4.64,4.53)physicians.GPT-4o and Claude 3.5 reached intermediate physician–level comprehensiveness(3.76±0.39,3.60±0.39)but junior-to-intermediate physician–level specificity(3.39±0.39,3.39±0.49).All LLMs scored higher on rare cases than intermediate physicians but fell below junior physicians in refractory-case specificity.AIassisted junior physicians showed marked gains in rare cases,with comprehensiveness rising from 2.32 to 4.29(84.8%),specificity from 2.24 to 4.26(90.8%),and readability from 2.76 to 4.59(66.0%),while specificity declined by 3.2%(3.17 to 3.07)in refractory cases.Error analysis showed complementary strengths,with physicians demonstrating reasoning stability and LLMs excelling in knowledge updating and risk management.Conclusions:LLMs performed variably in clinical decision-making tasks depending on case type,performing better in rare cases and worse in refractory cases requiring longitudinal reasoning.Complementary strengths between LLMs and physicians support case-and task-tailored human–AI collaboration.
基金supported by the Princess Nourah bint Abdulrahman University Researchers Supporting Project number(PNURSP2026R259)Princess Nourah bint Abdulrahman University,Riyadh,Saudi Arabia.Ashit Kumar Dutta would like to thank AlMaarefa University for supporting this research under project number MHIRSP2025017.
文摘Environmental problems are intensifying due to the rapid growth of the population,industry,and urban infrastructure.This expansion has resulted in increased air and water pollution,intensified urban heat island effects,and greater runoff from parks and other green spaces.Addressing these challenges requires prioritizing green infrastructure and other sustainable urban development strategies.This study introduces a novel Integrated Decision Support System that combines Pythagorean Fuzzy Sets with the Advanced Alternative Ranking Order Method allowing for Two-Step Normalization(AAROM-TN),enhanced by a dual weighting strategy.The weighting approach integrates the Criteria Importance Through Intercriteria Correlation(CRITIC)method with the Criteria Importance through Means and Standard Deviation(CIMAS)technique.The originality of the proposed framework lies in its ability to objectively quantify criteria importance using CRITIC,incorporate decision-makers’preferences through CIMAS,and capture the uncertainty and hesitation inherent in human judgment via Pythagorean Fuzzy Sets.A case study evaluating green infrastructure alternatives in metropolitan regions demonstrates the applicability and effectiveness of the framework.A sensitivity analysis is conducted to examine how variations in criteria weights affect the rankings and to evaluate the robustness of the results.Furthermore,a comparative analysis highlights the practical and financial implications of each alternative by assessing their respective strengths and weaknesses.
基金supported by the National Natural Science Foundation of China(U24B20183)the Pioneer Leading Goose+X Science and Technology Program of Zhejiang Province(2025C02018)。
文摘Dear Editor,This letter deals with the autonomous underwater vehicle(AUV)three dimensional(3D)trajectory tracking control chronically suffering from poor accuracy and efficiency in complex hydrodynamics.A state-of-the-art predictive adaptive controller(PAC)is proposed with a distinct dual closed-loop structure.
基金Supported by Changsha Tobacco Company Science and Technology Project(2020-2024A04).
文摘Starting from the foundational static traits underlying the growth and development of flue-cured tobacco, this research conducts a systematic examination of the phenomena and theoretical principles associated with environment-driven adaptive changes during its cultivation. It was found that environmental variables-including temperature, light, and moisture-elicit directional shifts in static traits ( e.g. , chemical composition, morphological architecture, and leaf tissue structure) toward enhanced environmental adaptation, characterized by graduality, juvenility, similarity, and correlativity. Upon alterations in ambient conditions, flue-cured tobacco modulates its static traits through integrated physical, chemical, and biological-genetic mechanisms, aiming to optimize resource utilization, mitigate environmental constraints, and preserve internal homeostasis alongside metabolic balance. The investigation further reveals that the adaptive scope of flue-cured tobacco to field environments is malleable and can be extended and elevated via adaptive conditioning commencing at the juvenile stage. In addition, the adaptive alignment between static traits and environmental parameters exerts a substantial impact on the plant s growth dynamics, yield performance, and quality attributes. Beyond its relevance to flue-cured tobacco, the proposed theory offers a meaningful framework for elucidating the pervasive adaptive strategies employed by plants and broader biological systems in response to environmental contingencies.
基金support of the National Key Research and Development Plan(No.2021YFB3302501)the financial support of the National Science Foundation of China(No.12161076)the financial support of the Fundamental Research Funds for the Central Universities(No.DUT25GF207).
文摘With the rapid development of artificial intelligence,intelligent air combat maneuver decision-making(ACMD)has garnered global attention.Although deep reinforcement learning provides a promising approach to ACMD,existing methods often suffer from rigid reward functions and limited adaptability to evolving adversarial strategies.Moreover,most research assumes open airspace,overlooking the influence of potential obstacles.In this paper,we address one-on-one within-visual-range ACMD in obstructed environments,and propose an improved Soft Actor-Critic(SAC)algorithm trained under a curriculum self-play framework.A maneuver strategy mirroring inference module is integrated to estimate each other's likely positions when visual obstruction occurs.By leveraging curriculum learning to guide progressive experience accumulation and self-play for adversarial evolution,our method enhances both training efficiency and tactical diversity.We further integrate an attention mechanism that dynamically adjusts the weights of sub-rewards,enabling the learned policy to adapt to rapidly changing air combat situations.Numerical simulations demonstrate that our enhanced SAC converges more quickly and achieves higher win rates than other baseline methods.An animation is available at bilibili.com/video/BV1BHVszHE98 for better illustration.
文摘In federated learning,backdoor attacks have become an important research topic with their wide application in processing sensitive datasets.Since federated learning detects or modifies local models through defense mechanisms during aggregation,it is difficult to conduct effective backdoor attacks.In addition,existing backdoor attack methods are faced with challenges,such as low backdoor accuracy,poor ability to evade anomaly detection,and unstable model training.To address these challenges,a method called adaptive simulation backdoor attack(ASBA)is proposed.Specifically,ASBA improves the stability of model training by manipulating the local training process and using an adaptive mechanism,the ability of the malicious model to evade anomaly detection by combing large simulation training and clipping,and the backdoor accuracy by introducing a stimulus model to amplify the impact of the backdoor in the global model.Extensive comparative experiments under five advanced defense scenarios show that ASBA can effectively evade anomaly detection and achieve high backdoor accuracy in the global model.Furthermore,it exhibits excellent stability and effectiveness after multiple rounds of attacks,outperforming state-of-the-art backdoor attack methods.
基金funded by scientific research projects under Grant JY2024B011.
文摘With the increasing complexity of industrial automation,planetary gearboxes play a vital role in largescale equipment transmission systems,directly impacting operational efficiency and safety.Traditional maintenance strategies often struggle to accurately predict the degradation process of equipment,leading to excessive maintenance costs or potential failure risks.However,existing prediction methods based on statistical models are difficult to adapt to nonlinear degradation processes.To address these challenges,this study proposes a novel condition-based maintenance framework for planetary gearboxes.A comprehensive full-lifecycle degradation experiment was conducted to collect raw vibration signals,which were then processed using a temporal convolutional network autoencoder with multi-scale perception capability to extract deep temporal degradation features,enabling the collaborative extraction of longperiod meshing frequencies and short-term impact features from the vibration signals.Kernel principal component analysis was employed to fuse and normalize these features,enhancing the characterization of degradation progression.A nonlinear Wiener process was used to model the degradation trajectory,with a threshold decay function introduced to dynamically adjust maintenance strategies,and model parameters optimized through maximum likelihood estimation.Meanwhile,the maintenance strategy was optimized to minimize costs per unit time,determining the optimal maintenance timing and preventive maintenance threshold.The comprehensive indicator of degradation trends extracted by this method reaches 0.756,which is 41.2%higher than that of traditional time-domain features;the dynamic threshold strategy reduces the maintenance cost per unit time to 55.56,which is 8.9%better than that of the static threshold optimization.Experimental results demonstrate significant reductions in maintenance costs while enhancing system reliability and safety.This study realizes the organic integration of deep learning and reliability theory in the maintenance of planetary gearboxes,provides an interpretable solution for the predictive maintenance of complex mechanical systems,and promotes the development of condition-based maintenance strategies for planetary gearboxes.
基金supported by the National Natural Science Foundation of China(Grant nos.42476264,41976224).
文摘Small RNAs(sRNAs)are important non-coding RNAs that usually play crucial roles in gene expression at the post-transcriptional level.The sRNAs have mostly been investigated in model microorganisms such as Escherichia coli and some pathogens.Nevertheless,microbial sRNAs from extreme environments such as the polar regions and deep sea have recently been discovered and analyzed for their unique roles in stress response,metabolic regulation and adaptation to extreme environments.These sRNAs fine-tune gene expression during oxidative and radiation stress,and modulate temperature and osmotic pressure responses.Representative sRNAs and their functions in thermophilic,psychrophilic,halophilic and radiation-tolerant bacteria are summarized in this review.Despite challenges in sample collection,RNA isolation,and functional annotation,the study of sRNAs in extreme environments provides opportunities for discovering novel regulatory mechanisms,applying them to biotechnology,and advancing our understanding of evolutionary adaptations.Looking ahead,high-throughput sequencing,synthetic biology,and multi-omics integration will bring new breakthroughs in discovering novel sRNAs and their functions and regulatory mechanisms.Such advancements are poised to enable comprehensive characterization of sRNA-mediated regulatory networks in extremophiles and unlock their biotechnological potential through mechanism-driven applications.
基金support from the National Natural Science Foundation of China(Nos.22293011,T2341001)the Major Science and Technology Project of Anhui Province(202203a06020010).
文摘Epoxy resins are widely employed in wind turbine blades,drone rotors,and automotive interiors due to their excel-lent mechani-cal proper-ties and long service life.However,their insoluble and infusible cross-linked networks pose a significant re-cycling challenge,particularly with the impending retirement of the first generation of wind turbine blades.In this work,we reported a fully bio-based epoxy Vitrimer(FEP)incorporat-ing a dual-dynamic covalent network design and systematically investigated the influence of the 1,5,7-triazabicyclo[4.4.0]dec-5-ene(TBD)catalyst on its curing kinetics,thermal/mechan-ical properties,dynamic exchange behavior,and degradation performance in a mild alkaline solution.Compared to conventional epoxy resins,FEP exhibited superior tensile strength and elongation at break at an optimal TBD concentration(2 wt%),achieving an excellent strength-toughness balance.The presence of TBD accelerated the exchange rates of both disulfide and ester bonds,endowing FEP with notable stress relaxation at elevated tempera-tures.Moreover,FEP demonstrated complete dissolution in 1 mol/L NaOH within 6 h at 25℃.These results underscored the exceptional strength,toughness,and recyclability of FEP,positioning it as a promising,environmentally friendly matrix resin for next-generation appli-cations in the new energy sector.
基金supported by Singapore RIE2025 Manufacturing,Trade and Connectivity Industry Alignment Fund-Pre-Positioning(IAF-PP)under Grant M24N2a0039 through WP2-Intelligent Switching Controlthe National Research Foundation Singapore under its AI Singapore Programme under Grant AISG4-GC-2023-007-1B.
文摘In this paper,we propose a fuzzy logic-based coded event-triggered control with self-adjustable prescribed performance(FL-CEC-SPP)to address the trade-off between control performance and communication efficiency in resource-constrained networked control systems.The method integrates a fuzzy-coded event-triggered controller into a coded control framework to dynamically adjust the triggering threshold,thereby reducing unnecessary transmissions while maintaining system stability.A self-adjustable prescribed performance constraint is also incorporated to ensure that the tracking error remains within predefined bounds under arbitrary initial conditions.Theoretical analyses and simulation comparisons show that the method proposed in this paper maintains good tracking performance and stability while reducing the communication burden,and has wide applications in resource-constrained network control systems.
基金funded by the Natural Science Foundation of Chongqing Municipality,grant number CSTB2022NSCQ-MSX0503.
文摘Gait recognition is a key biometric for long-distance identification,yet its performance is severely degraded by real-world challenges such as varying clothing,carrying conditions,and changing viewpoints.While combining silhouette and skeleton data is a promising direction,effectively fusing these heterogeneous modalities and adaptively weighting their contributions in response to diverse conditions remains a central problem.This paper introduces GaitMAFF,a novelMulti-modal Adaptive Feature Fusion Network,to address this challenge.Our approach first transforms discrete skeleton joints into a dense SkeletonMap representation to align with silhouettes,then employs an attention-based module to dynamically learn the fusion weights between the two modalities.These fused features are processed by a powerful spatio-temporal backbone withWeighted Global-Local Feature FusionModules(WFFM)to learn a discriminative representation.Extensive experiments on the challenging CCPG and Gait3D datasets show that GaitMAFF achieves state-of-the-art performance,with an average Rank-1 accuracy of 84.6%on CCPG and 58.7%on Gait3D.These results demonstrate that our adaptive fusion strategy effectively integrates complementary multimodal information,significantly enhancing gait recognition robustness and accuracy in complex scenes and providing a practical solution for real-world applications.
基金supported by the National Natural Science Foundation of China(NSFC)under Grant number:82171965.
文摘Designing appropriate loss functions is critical to the success of supervised learning models.However,most conventional losses are fixed and manually designed,making them suboptimal for diverse and dynamic learning scenarios.In this work,we propose an Adaptive Meta-Loss Network(Adaptive-MLN)that learns to generate taskagnostic loss functions tailored to evolving classification problems.Unlike traditional methods that rely on static objectives,Adaptive-MLN treats the loss function itself as a trainable component,parameterized by a shallow neural network.To enable flexible,gradient-free optimization,we introduce a hybrid evolutionary approach that combines GeneticAlgorithms(GA)for global exploration and Evolution Strategies(ES)for local refinement.This co-evolutionary process dynamically adjusts the loss landscape,improvingmodel generalization without relying on analytic gradients or handcrafted heuristics.Experimental evaluations on synthetic tasks and the CIFAR-10 andMNIST datasets demonstrate that our approach consistently outperforms standard losses such as Cross-Entropy and Mean Squared Error in terms of accuracy,convergence,and adaptability.
基金funded by the National Natural Science Foundation of China(NSFC)under Grant Number 72071209.
文摘Modern battlefields exhibit high dynamism,where traditional static weighting methods in combat effectiveness assessment fail to capture real-time changes in indicator values,leading to limited assessment accuracy—especially critical in scenarios like sudden electronic warfare or degraded command,where static weights cannot reflect the operational value decay or surge of key indicators.To address this issue,this study proposes a dynamic adaptive weightingmethod for evaluation indicators based onG1-CRITIC-PIVW.First,theG1(Sequential Relationship Analysis Method)subjective weighting method—translates expert knowledge into indicator importance rankings—leverages expert knowledge to quantify the relative importance of indicators via sequential relationship ranking,while the CRITIC(Criteria Importance Through Intercriteria Correlation)objective weighting method—derives weights from data characteristics by integrating variability and inter-correlations—calculates weights by integrating indicator variability and inter-indicator correlations,ensuring data-driven objectivity.These two sets of weights are then fused using a deviation coefficient optimization model,minimizing the squared deviation from a reference weight and adjusting the fusion coefficient via Spearman’s rank correlation to resolve potential conflicts between subjective and objective judgments.Subsequently,the PIVW(Punishment-Incentive VariableWeight)theory—adapts weights to realtime indicator performance via penalty/incentive rules—is applied for dynamic adjustment.Scenario-specific penalty λ_(1) and incentive λ_(2) thresholds are set based on operational priorities and indicator volatility,penalizing indicators with values below λ_(1) and incentivizing those exceeding λ_(2) to reflect real-time indicator performance.Experimental validation was conducted using an Air Defense and Anti-Missile(ADAM)system effectiveness assessment framework,with data covering 7 indicators across 3 combat scenarios.Results show that compared to static weighting methods,the proposed method reduces MAE(Mean Absolute Error)by 15%-20% and weighted decision error rate by 84.2%,effectively reducing overestimation/underestimation of combat effectiveness in dynamic scenarios;compared to Entropy-TOPSIS,it lowers MAE by 12% while achieving a weighted Kendall’sτconsistency coefficient of 0.85,ensuring higher alignment with expert judgment.This method enhances the accuracy and scenario adaptability of effectiveness assessment,providing reliable decision support for dynamic battlefield environments.
文摘The rapidly evolving cybersecurity threat landscape exposes a critical flaw in traditional educational programs where static curricula cannot adapt swiftly to novel attack vectors.This creates a significant gap between theoretical knowledge and the practical defensive capabilities needed in the field.To address this,we propose TeachSecure-CTI,a novel framework for adaptive cybersecurity curriculumgeneration that integrates real-time Cyber Threat Intelligence(CTI)with AI-driven personalization.Our framework employs a layered architecture featuring a CTI ingestion and clusteringmodule,natural language processing for semantic concept extraction,and a reinforcement learning agent for adaptive content sequencing.Bydynamically aligning learningmaterialswithboththe evolving threat environment and individual learner profiles,TeachSecure-CTI ensures content remains current,relevant,and tailored.A 12-week study with 150 students across three institutions demonstrated that the framework improves learning gains by 34%,significantly exceeding the 12%–21%reported in recent literature.The system achieved 84.8%personalization accuracy,85.9%recognition accuracy for MITRE ATT&CK tactics,and a 31%faster competency development rate compared to static curricula.These findings have implications beyond academia,extending to workforce development,cyber range training,and certification programs.By bridging the gap between dynamic threats and static educational materials,TeachSecure-CTI offers an empirically validated,scalable solution for cultivating cybersecurity professionals capable of responding to modern threats.
基金funding from the National Key Research and Development Program of China(No.2018YFE0110000)the National Natural Science Foundation of China(No.11274259,No.11574258)the Science and Technology Commission Foundation of Shanghai(21DZ1205500)in support of the present research.
文摘While reinforcement learning-based underwater acoustic adaptive modulation shows promise for enabling environment-adaptive communication as supported by extensive simulation-based research,its practical performance remains underexplored in field investigations.To evaluate the practical applicability of this emerging technique in adverse shallow sea channels,a field experiment was conducted using three communication modes:orthogonal frequency division multiplexing(OFDM),M-ary frequency-shift keying(MFSK),and direct sequence spread spectrum(DSSS)for reinforcement learning-driven adaptive modulation.Specifically,a Q-learning method is used to select the optimal modulation mode according to the channel quality quantified by signal-to-noise ratio,multipath spread length,and Doppler frequency offset.Experimental results demonstrate that the reinforcement learning-based adaptive modulation scheme outperformed fixed threshold detection in terms of total throughput and average bit error rate,surpassing conventional adaptive modulation strategies.
基金supported in part by the National Natural Science Foundation of China(62173208)Taishan Scholar Project of Shandong Province of China(tsqn202103061)the National Science and Technology Council(NSTC),Taiwan,China(NSTC 113-2221-E-006-145-MY2)。
文摘This paper explores the adaptive exponentially designated-time stabilization issue via event-triggered feedback for a kind of uncertain high-order nonlinear systems.The motivation mainly comes from the following two challenges:the undesired singularity problem arising from infinite control gains at the prescribed-time instant,the effective trade-off between the control amplitude and the triggering duration.The goal is to build an event-triggered mechanism comprising a skillful triggered rule alongside a time-dependent threshold.Utilizing the designed control strategy,the solutions'existence and the prevention of Zeno phenomenon are successfully guaranteed by using a new transformation equipped with a time-varying function and redesigning the continuous state-feedback dominance approach with an array of integral functions involving embedded sign functions.Better than existing prescribed-time methods,our approach not only ensures that state variables converge to a small compact set before a designated time and stay there henceforth,and converge to the origin exponentially,but also ensures that the controller continuously works on the whole-time horizon.Two illustrative examples are given to show the effectiveness of the devised scheme.
基金supported by the Funds for Central-Guided Local Science and Technology Development(Grant No.202407AC110005)Key Technologies for the Construction of a Whole-Process Intelligent Service System for Neuroendocrine Neoplasm.Supported by 2023 Opening Research Fund of Yunnan Key Laboratory of Digital Communications(YNJTKFB-20230686,YNKLDC-KFKT-202304).
文摘In image analysis,high-precision semantic segmentation predominantly relies on supervised learning.Despite significant advancements driven by deep learning techniques,challenges such as class imbalance and dynamic performance evaluation persist.Traditional weighting methods,often based on pre-statistical class counting,tend to overemphasize certain classes while neglecting others,particularly rare sample categories.Approaches like focal loss and other rare-sample segmentation techniques introduce multiple hyperparameters that require manual tuning,leading to increased experimental costs due to their instability.This paper proposes a novel CAWASeg framework to address these limitations.Our approach leverages Grad-CAM technology to generate class activation maps,identifying key feature regions that the model focuses on during decision-making.We introduce a Comprehensive Segmentation Performance Score(CSPS)to dynamically evaluate model performance by converting these activation maps into pseudo mask and comparing them with Ground Truth.Additionally,we design two adaptive weights for each class:a Basic Weight(BW)and a Ratio Weight(RW),which the model adjusts during training based on real-time feedback.Extensive experiments on the COCO-Stuff,CityScapes,and ADE20k datasets demonstrate that our CAWASeg framework significantly improves segmentation performance for rare sample categories while enhancing overall segmentation accuracy.The proposed method offers a robust and efficient solution for addressing class imbalance in semantic segmentation tasks.
