Background:With the global popularity of short videos,particularly among young people,short video addiction has become a worrying phenomenon that poses significant risks to individual health and adaptability.Self-regu...Background:With the global popularity of short videos,particularly among young people,short video addiction has become a worrying phenomenon that poses significant risks to individual health and adaptability.Self-regulated learning(SRL)strategies are key factors in predicting learning outcomes.This study,based on the SRL theory,uses short video addiction as the independent variable,SRL strategies as the mediating variable,and learning well-being as the outcome variable,aiming to reveal the relationships among short video addiction,self-regulated learning,and learning well-being among Chinese college students.Methods:Using a cross-sectional study design and applying the snowball sampling technique,an online survey was administered to Chinese undergraduate students.A total of 706 valid questionnaires were collected,with an effective response rate of 85.7%.The average age of the participants was 20.5 years.Results:The results of structural equation modeling indicate that 7 hypotheses were supported.Short video addiction was negatively correlated with self-regulated learning strategies(preparatory,performance,and appraisal strategy),while SRL strategies were positively correlated with learning well-being.Additionally,short video addiction had a mediating effect on learning well-being through the three types of SRL strategies.The three types of SRL strategies explained 39%of the variance in learning well-being.Conclusion:Previous research has typically focused on the impact of self-regulated learning strategies on media addiction or problematic media use.This study,based on the SRL model,highlights the negative issues caused by short video addiction and emphasizes the importance of cultivating self-regulation abilities and media literacy.Short video addiction stems from failures in trait self-regulation,which naturally impairs the ability to effectively engage in self-regulation during the learning process.This study confirms and underscores that the SRL model can serve as an effective theoretical framework for helping students prevent short video addiction,engage in high-quality learning,and consequently enhance their learning well-being.展开更多
Model for spoken is expected to overcome difficulties in teaching and learning Indonesian language for foreign speakers. Language anxiety is the anxiety that arises when a person learns foreign language. Foreign Langu...Model for spoken is expected to overcome difficulties in teaching and learning Indonesian language for foreign speakers. Language anxiety is the anxiety that arises when a person learns foreign language. Foreign Language Anxiety （anxiety to learn a foreign language） is of concern or negative emotional reactions that arise when studying or using foreign language. Self-regulated learning is an active and constructive process undertaken by learners in setting goals for their learning and trying to monitor, regulate, and control of cognition, motivation, and behavior, then everything is directed and driven by purpose and adapted to the context and environment. The research method used is an R and D （research and development） method with a sample of foreign speakers of Chinese. Variables that receive interference are the ability to speak in Indonesian, while the variables used to interfere with the self-regulated learning and language anxiety as a variable controller. Intrapersonal factors become barriers that cause stuttering speech limited due to the mastering subject content. On the basis of that, this speaking model applies the principle of self-regulated learning in the learning process, using a communicative and contextual approach. This model intended for foreign speakers who learn Indonesian language outside of Indonesia, so to bring the atmosphere mandated in sociolinguistic built through media and relevant teaching methods.展开更多
This study explored the nature and use of technology-based self-regulated learning(SRL)strategies among the Chinese university students.A total of 20 undergraduate students in China's Mainland were invited to part...This study explored the nature and use of technology-based self-regulated learning(SRL)strategies among the Chinese university students.A total of 20 undergraduate students in China's Mainland were invited to participate in a focus group interview.The students reported using four types of technology-based SRL strategies including cognitive,meta-cognitive,social behavioral,and motivational regulation strategies.Among the strategies,technology-based vocabulary learning was reported to be a dominant strategy by the students.This study opens a new window to understanding how English as a foreign language(EFL)students utilize different strategies to learn English in technology-based learning context.展开更多
The purpose of this paper is to argue the effectiveness of self-regulated learning in English education in Chinese college classroom instruction. A study is given to show whether the introduction of self-regulated lea...The purpose of this paper is to argue the effectiveness of self-regulated learning in English education in Chinese college classroom instruction. A study is given to show whether the introduction of self-regulated learning can help improve Chinese college students' English learning, and help them perform better in the National English test-CET-4 (College English Test Level-4,).展开更多
The present study adopts a mixed method design to investigate the effect of seven classroom assessment(CA)features on student self-regulated learning(SRL)and further explored factors that influenced the effect.Twelve ...The present study adopts a mixed method design to investigate the effect of seven classroom assessment(CA)features on student self-regulated learning(SRL)and further explored factors that influenced the effect.Twelve teachers and their students(valid data points tallying 630)from three Chinese high schools participated in the study.Structural equational modelling results showed that the CA features had varied impacts.Specifically,self-assessment most effectively enhanced SRL,followed by teacher instruction and structured guidance,then teacher feedback;assessment task and student choice had mixed impacts.Peer-assessment and CA environment reduced SRL.Five influencing factors were revealed through both teacher and student interviews,namely,student engagement with the assessment task,student dependence on authority,prospective gains in the gaokao,intractable motivation and learning approach,and student-teacher relationship.The research has practical implications for SRL promotion.展开更多
The study investigated the effects of an intervention program on self-regulated learning designed for second language learners. One hundred and twenty participants who were sophomore English majors at a university in ...The study investigated the effects of an intervention program on self-regulated learning designed for second language learners. One hundred and twenty participants who were sophomore English majors at a university in China were randomly assigned to either the treatment or the control group. The intervention was composed of six weekly two-hour training sessions that focus on five main variables of self-regulatory processes： goal setting, self-efficacy, time and study environment management, language learning strategies, and attribution. The evaluation of the effectiveness of the intervention included mukiple outcome variables, which were grouped into three categories： students＇ motivational beliefs, students＇ strategy use, and students＇ academic performance. The results of the immediate training effects on goal setting, self-efficacy, attribution, time and study environment management, memory strategy, compensation strategy, metacognitive strategy and second language proficiency confirmed that academic self-regulation is a trainable student characteristic and self-regulation training can be used effectively in a second language classroom setting. The feature of the current study design allows for systematically examining and evaluating both motivational variables and learning strategies in the context of second language learning.展开更多
The present study is set in the context of ongoing educational reform that advocates fostering self-regulated learners and using assessment to improve learning.Drawing on existent research on classroom assessment(CA)a...The present study is set in the context of ongoing educational reform that advocates fostering self-regulated learners and using assessment to improve learning.Drawing on existent research on classroom assessment(CA)and self-regulated learning(SRL),the authors have formulated a conceptual framework outlining the CA features that promote SRL among students.Guided by this framework,the 12 high school teachers’CA practice was scrutinized to find out to what extent their CA was pro-SRL.Based on interview data and classroom observation,gaps were found in Chinese high school teachers’CA.First,CA tasks are primarily low-level closed-end problems,with rare exceptions.Second,students are not allowed much autonomy in CA.Third,self-assessment practice is mostly self-grading.Fourth,peer-assessment is uncommon and mainly involves simply marking peers’work.Fifth,teacher feedback is focused on task and process levels;regulation-level feedback is less common.Sixth,despite teachers’encouragement,most students feel threatened by CA.展开更多
In this paper,I use an autoethnographical approach,coupled with existing research literature on Chinese learners and learning,to reflect upon my own experiences as a junior high school student in order to explore how ...In this paper,I use an autoethnographical approach,coupled with existing research literature on Chinese learners and learning,to reflect upon my own experiences as a junior high school student in order to explore how Chinese students perceive their learning,and how they establish and justify their own sense of self-regulation in learning.It is found there is a hybrid of nuanced cultural meanings underneath the self-regulated learning experiences in the Chinese context.展开更多
Cyclohexene is an important raw material in the production of nylon.Selective hydrogenation of benzene is a key method for preparing cyclohexene.However,the Ru catalysts used in current industrial processes still face...Cyclohexene is an important raw material in the production of nylon.Selective hydrogenation of benzene is a key method for preparing cyclohexene.However,the Ru catalysts used in current industrial processes still face challenges,including high metal usage,high process costs,and low cyclohexene yield.This study utilizes existing literature data combined with machine learning methods to analyze the factors influencing benzene conversion,cyclohexene selectivity,and yield in the benzene hydrogenation to cyclohexene reaction.It constructs predictive models based on XGBoost and Random Forest algorithms.After analysis,it was found that reaction time,Ru content,and space velocity are key factors influencing cyclohexene yield,selectivity,and benzene conversion.Shapley Additive Explanations(SHAP)analysis and feature importance analysis further revealed the contribution of each variable to the reaction outcomes.Additionally,we randomly generated one million variable combinations using the Dirichlet distribution to attempt to predict high-yield catalyst formulations.This paper provides new insights into the application of machine learning in heterogeneous catalysis and offers some reference for further research.展开更多
The uplift resistance of the soil overlying shield tunnels significantly impacts their anti-floating stability.However,research on uplift resistance concerning special-shaped shield tunnels is limited.This study combi...The uplift resistance of the soil overlying shield tunnels significantly impacts their anti-floating stability.However,research on uplift resistance concerning special-shaped shield tunnels is limited.This study combines numerical simulation with machine learning techniques to explore this issue.It presents a summary of special-shaped tunnel geometries and introduces a shape coefficient.Through the finite element software,Plaxis3D,the study simulates six key parameters—shape coefficient,burial depth ratio,tunnel’s longest horizontal length,internal friction angle,cohesion,and soil submerged bulk density—that impact uplift resistance across different conditions.Employing XGBoost and ANN methods,the feature importance of each parameter was analyzed based on the numerical simulation results.The findings demonstrate that a tunnel shape more closely resembling a circle leads to reduced uplift resistance in the overlying soil,whereas other parameters exhibit the contrary effects.Furthermore,the study reveals a diminishing trend in the feature importance of buried depth ratio,internal friction angle,tunnel longest horizontal length,cohesion,soil submerged bulk density,and shape coefficient in influencing uplift resistance.展开更多
Federated Learning(FL)has become a leading decentralized solution that enables multiple clients to train a model in a collaborative environment without directly sharing raw data,making it suitable for privacy-sensitiv...Federated Learning(FL)has become a leading decentralized solution that enables multiple clients to train a model in a collaborative environment without directly sharing raw data,making it suitable for privacy-sensitive applications such as healthcare,finance,and smart systems.As the field continues to evolve,the research field has become more complex and scattered,covering different system designs,training methods,and privacy techniques.This survey is organized around the three core challenges:how the data is distributed,how models are synchronized,and how to defend against attacks.It provides a structured and up-to-date review of FL research from 2023 to 2025,offering a unified taxonomy that categorizes works by data distribution(Horizontal FL,Vertical FL,Federated Transfer Learning,and Personalized FL),training synchronization(synchronous and asynchronous FL),optimization strategies,and threat models(data leakage and poisoning attacks).In particular,we summarize the latest contributions in Vertical FL frameworks for secure multi-party learning,communication-efficient Horizontal FL,and domain-adaptive Federated Transfer Learning.Furthermore,we examine synchronization techniques addressing system heterogeneity,including straggler mitigation in synchronous FL and staleness management in asynchronous FL.The survey covers security threats in FL,such as gradient inversion,membership inference,and poisoning attacks,as well as their defense strategies that include privacy-preserving aggregation and anomaly detection.The paper concludes by outlining unresolved issues and highlighting challenges in handling personalized models,scalability,and real-world adoption.展开更多
Split Learning(SL)has been promoted as a promising collaborative machine learning technique designed to address data privacy and resource efficiency.Specifically,neural networks are divided into client and server subn...Split Learning(SL)has been promoted as a promising collaborative machine learning technique designed to address data privacy and resource efficiency.Specifically,neural networks are divided into client and server subnetworks in order to mitigate the exposure of sensitive data and reduce the overhead on client devices,thereby making SL particularly suitable for resource-constrained devices.Although SL prevents the direct transmission of raw data,it does not alleviate entirely the risk of privacy breaches.In fact,the data intermediately transmitted to the server sub-model may include patterns or information that could reveal sensitive data.Moreover,achieving a balance between model utility and data privacy has emerged as a challenging problem.In this article,we propose a novel defense approach that combines:(i)Adversarial learning,and(ii)Network channel pruning.In particular,the proposed adversarial learning approach is specifically designed to reduce the risk of private data exposure while maintaining high performance for the utility task.On the other hand,the suggested channel pruning enables the model to adaptively adjust and reactivate pruned channels while conducting adversarial training.The integration of these two techniques reduces the informativeness of the intermediate data transmitted by the client sub-model,thereby enhancing its robustness against attribute inference attacks without adding significant computational overhead,making it wellsuited for IoT devices,mobile platforms,and Internet of Vehicles(IoV)scenarios.The proposed defense approach was evaluated using EfficientNet-B0,a widely adopted compact model,along with three benchmark datasets.The obtained results showcased its superior defense capability against attribute inference attacks compared to existing state-of-the-art methods.This research’s findings demonstrated the effectiveness of the proposed channel pruning-based adversarial training approach in achieving the intended compromise between utility and privacy within SL frameworks.In fact,the classification accuracy attained by the attackers witnessed a drastic decrease of 70%.展开更多
Knowledge distillation has become a standard technique for compressing large language models into efficient student models,but existing methods often struggle to balance prediction accuracy with explanation quality.Re...Knowledge distillation has become a standard technique for compressing large language models into efficient student models,but existing methods often struggle to balance prediction accuracy with explanation quality.Recent approaches such as Distilling Step-by-Step(DSbS)introduce explanation supervision,yet they apply it in a uniform manner that may not fully exploit the different learning dynamics of prediction and explanation.In this work,we propose a task-structured curriculum learning(TSCL)framework that structures training into three sequential phases:(i)prediction-only,to establish stable feature representations;(ii)joint prediction-explanation,to align task outputs with rationale generation;and(iii)explanation-only,to refine the quality of rationales.This design provides a simple but effective modification to DSbS,requiring no architectural changes and adding negligible training cost.We justify the phase scheduling with ablation studies and convergence analysis,showing that an initial prediction-heavy stage followed by a balanced joint phase improves both stability and explanation alignment.Extensive experiments on five datasets(e-SNLI,ANLI,CommonsenseQA,SVAMP,and MedNLI)demonstrate that TSCL consistently outperforms strong baselines,achieving gains of+1.7-2.6 points in accuracy and 0.8-1.2 in ROUGE-L,corresponding to relative error reductions of up to 21%.Beyond lexical metrics,human evaluation and ERASERstyle faithfulness diagnostics confirm that TSCL produces more faithful and informative explanations.Comparative training curves further reveal faster convergence and lower variance across seeds.Efficiency analysis shows less than 3%overhead in wall-clock training time and no additional inference cost,making the approach practical for realworld deployment.This study demonstrates that a simple task-structured curriculum can significantly improve the effectiveness of knowledge distillation.By separating and sequencing objectives,TSCL achieves a better balance between accuracy,stability,and explanation quality.The framework generalizes across domains,including medical NLI,and offers a principled recipe for future applications in multimodal reasoning and reinforcement learning.展开更多
Latest digital advancements have intensified the necessity for adaptive,data-driven and socially-centered learning ecosystems.This paper presents the formulation of a cross-platform,innovative,gamified and personalize...Latest digital advancements have intensified the necessity for adaptive,data-driven and socially-centered learning ecosystems.This paper presents the formulation of a cross-platform,innovative,gamified and personalized Learning Ecosystem,which integrates 3D/VR environments,as well as machine learning algorithms,and business intelligence frameworks to enhance learner-centered education and inferenced decision-making.This Learning System makes use of immersive,analytically assessed virtual learning spaces,therefore facilitating real-time monitoring of not just learning performance,but also overall engagement and behavioral patterns,via a comprehensive set of sustainability-oriented ESG-aligned Key Performance Indicators(KPIs).Machine learning models support predictive analysis,personalized feedback,and hybrid recommendation mechanisms,whilst dedicated dashboards translate complex educational data into actionable insights for all Use Cases of the System(Educational Institutions,Educators and Learners).Additionally,the presented Learning System introduces a structured Mentoring and Consulting Subsystem,thence reinforcing human-centered guidance alongside automated intelligence.The Platform’s modular architecture and simulation-centered evaluation approach actively support personalized,and continuously optimized learning pathways.Thence,it exemplifies a mature,adaptive Learning Ecosystem,supporting immersive technologies,analytics,and pedagogical support,hence,contributing to contemporary digital learning innovation and sociotechnical transformation in education.展开更多
Automated grading of dandruff severity is a clinically significant but challenging task due to the inherent ordinal nature of severity levels and the high prevalence of label noise from subjective expert annotations.S...Automated grading of dandruff severity is a clinically significant but challenging task due to the inherent ordinal nature of severity levels and the high prevalence of label noise from subjective expert annotations.Standard classification methods fail to address these dual challenges,limiting their real-world performance.In this paper,a novel,three-phase training framework is proposed that learns a robust ordinal classifier directly from noisy labels.The approach synergistically combines a rank-based ordinal regression backbone with a cooperative,semi-supervised learning strategy to dynamically partition the data into clean and noisy subsets.A hybrid training objective is then employed,applying a supervised ordinal loss to the clean set.The noisy set is simultaneously trained using a dualobjective that combines a semi-supervised ordinal loss with a parallel,label-agnostic contrastive loss.This design allows themodel to learn fromthe entire noisy subset while using contrastive learning to mitigate the risk of error propagation frompotentially corrupt supervision.Extensive experiments on a new,large-scale,multi-site clinical dataset validate our approach.Themethod achieves state-of-the-art performance with 80.71%accuracy and a 76.86%F1-score,significantly outperforming existing approaches,including a 2.26%improvement over the strongest baseline method.This work provides not only a robust solution for a practical medical imaging problem but also a generalizable framework for other tasks plagued by noisy ordinal labels.展开更多
基金supported by Fundamental Research Funds for the Central Universities in China(Grant Number:2022NTSS52)First-Class Education Discipline Development of Beijing Normal University:Excellence Action Project(Grant Number:YLXKPYXSDW202408)Beijing Institute of Education 2024 Youth Research Projects“Research on the Transformation of Training Aimed at Improving the Work of School Principals in the Capital”(Grant Number:QZ2024-02).
文摘Background:With the global popularity of short videos,particularly among young people,short video addiction has become a worrying phenomenon that poses significant risks to individual health and adaptability.Self-regulated learning(SRL)strategies are key factors in predicting learning outcomes.This study,based on the SRL theory,uses short video addiction as the independent variable,SRL strategies as the mediating variable,and learning well-being as the outcome variable,aiming to reveal the relationships among short video addiction,self-regulated learning,and learning well-being among Chinese college students.Methods:Using a cross-sectional study design and applying the snowball sampling technique,an online survey was administered to Chinese undergraduate students.A total of 706 valid questionnaires were collected,with an effective response rate of 85.7%.The average age of the participants was 20.5 years.Results:The results of structural equation modeling indicate that 7 hypotheses were supported.Short video addiction was negatively correlated with self-regulated learning strategies(preparatory,performance,and appraisal strategy),while SRL strategies were positively correlated with learning well-being.Additionally,short video addiction had a mediating effect on learning well-being through the three types of SRL strategies.The three types of SRL strategies explained 39%of the variance in learning well-being.Conclusion:Previous research has typically focused on the impact of self-regulated learning strategies on media addiction or problematic media use.This study,based on the SRL model,highlights the negative issues caused by short video addiction and emphasizes the importance of cultivating self-regulation abilities and media literacy.Short video addiction stems from failures in trait self-regulation,which naturally impairs the ability to effectively engage in self-regulation during the learning process.This study confirms and underscores that the SRL model can serve as an effective theoretical framework for helping students prevent short video addiction,engage in high-quality learning,and consequently enhance their learning well-being.
文摘Model for spoken is expected to overcome difficulties in teaching and learning Indonesian language for foreign speakers. Language anxiety is the anxiety that arises when a person learns foreign language. Foreign Language Anxiety （anxiety to learn a foreign language） is of concern or negative emotional reactions that arise when studying or using foreign language. Self-regulated learning is an active and constructive process undertaken by learners in setting goals for their learning and trying to monitor, regulate, and control of cognition, motivation, and behavior, then everything is directed and driven by purpose and adapted to the context and environment. The research method used is an R and D （research and development） method with a sample of foreign speakers of Chinese. Variables that receive interference are the ability to speak in Indonesian, while the variables used to interfere with the self-regulated learning and language anxiety as a variable controller. Intrapersonal factors become barriers that cause stuttering speech limited due to the mastering subject content. On the basis of that, this speaking model applies the principle of self-regulated learning in the learning process, using a communicative and contextual approach. This model intended for foreign speakers who learn Indonesian language outside of Indonesia, so to bring the atmosphere mandated in sociolinguistic built through media and relevant teaching methods.
文摘This study explored the nature and use of technology-based self-regulated learning(SRL)strategies among the Chinese university students.A total of 20 undergraduate students in China's Mainland were invited to participate in a focus group interview.The students reported using four types of technology-based SRL strategies including cognitive,meta-cognitive,social behavioral,and motivational regulation strategies.Among the strategies,technology-based vocabulary learning was reported to be a dominant strategy by the students.This study opens a new window to understanding how English as a foreign language(EFL)students utilize different strategies to learn English in technology-based learning context.
文摘The purpose of this paper is to argue the effectiveness of self-regulated learning in English education in Chinese college classroom instruction. A study is given to show whether the introduction of self-regulated learning can help improve Chinese college students' English learning, and help them perform better in the National English test-CET-4 (College English Test Level-4,).
文摘The present study adopts a mixed method design to investigate the effect of seven classroom assessment(CA)features on student self-regulated learning(SRL)and further explored factors that influenced the effect.Twelve teachers and their students(valid data points tallying 630)from three Chinese high schools participated in the study.Structural equational modelling results showed that the CA features had varied impacts.Specifically,self-assessment most effectively enhanced SRL,followed by teacher instruction and structured guidance,then teacher feedback;assessment task and student choice had mixed impacts.Peer-assessment and CA environment reduced SRL.Five influencing factors were revealed through both teacher and student interviews,namely,student engagement with the assessment task,student dependence on authority,prospective gains in the gaokao,intractable motivation and learning approach,and student-teacher relationship.The research has practical implications for SRL promotion.
文摘The study investigated the effects of an intervention program on self-regulated learning designed for second language learners. One hundred and twenty participants who were sophomore English majors at a university in China were randomly assigned to either the treatment or the control group. The intervention was composed of six weekly two-hour training sessions that focus on five main variables of self-regulatory processes： goal setting, self-efficacy, time and study environment management, language learning strategies, and attribution. The evaluation of the effectiveness of the intervention included mukiple outcome variables, which were grouped into three categories： students＇ motivational beliefs, students＇ strategy use, and students＇ academic performance. The results of the immediate training effects on goal setting, self-efficacy, attribution, time and study environment management, memory strategy, compensation strategy, metacognitive strategy and second language proficiency confirmed that academic self-regulation is a trainable student characteristic and self-regulation training can be used effectively in a second language classroom setting. The feature of the current study design allows for systematically examining and evaluating both motivational variables and learning strategies in the context of second language learning.
文摘The present study is set in the context of ongoing educational reform that advocates fostering self-regulated learners and using assessment to improve learning.Drawing on existent research on classroom assessment(CA)and self-regulated learning(SRL),the authors have formulated a conceptual framework outlining the CA features that promote SRL among students.Guided by this framework,the 12 high school teachers’CA practice was scrutinized to find out to what extent their CA was pro-SRL.Based on interview data and classroom observation,gaps were found in Chinese high school teachers’CA.First,CA tasks are primarily low-level closed-end problems,with rare exceptions.Second,students are not allowed much autonomy in CA.Third,self-assessment practice is mostly self-grading.Fourth,peer-assessment is uncommon and mainly involves simply marking peers’work.Fifth,teacher feedback is focused on task and process levels;regulation-level feedback is less common.Sixth,despite teachers’encouragement,most students feel threatened by CA.
文摘In this paper,I use an autoethnographical approach,coupled with existing research literature on Chinese learners and learning,to reflect upon my own experiences as a junior high school student in order to explore how Chinese students perceive their learning,and how they establish and justify their own sense of self-regulation in learning.It is found there is a hybrid of nuanced cultural meanings underneath the self-regulated learning experiences in the Chinese context.
基金Supported by CAS Basic and Interdisciplinary Frontier Scientific Research Pilot Project(XDB1190300,XDB1190302)Youth Innovation Promotion Association CAS(Y2021056)+1 种基金Joint Fund of the Yulin University and the Dalian National Laboratory for Clean Energy(YLU-DNL Fund 2022007)The special fund for Science and Technology Innovation Teams of Shanxi Province(202304051001007)。
文摘Cyclohexene is an important raw material in the production of nylon.Selective hydrogenation of benzene is a key method for preparing cyclohexene.However,the Ru catalysts used in current industrial processes still face challenges,including high metal usage,high process costs,and low cyclohexene yield.This study utilizes existing literature data combined with machine learning methods to analyze the factors influencing benzene conversion,cyclohexene selectivity,and yield in the benzene hydrogenation to cyclohexene reaction.It constructs predictive models based on XGBoost and Random Forest algorithms.After analysis,it was found that reaction time,Ru content,and space velocity are key factors influencing cyclohexene yield,selectivity,and benzene conversion.Shapley Additive Explanations(SHAP)analysis and feature importance analysis further revealed the contribution of each variable to the reaction outcomes.Additionally,we randomly generated one million variable combinations using the Dirichlet distribution to attempt to predict high-yield catalyst formulations.This paper provides new insights into the application of machine learning in heterogeneous catalysis and offers some reference for further research.
基金Guangzhou Metro Scientific Research Project(No.JT204-100111-23001)Chongqing Municipal Special Project for Technological Innovation and Application Development(No.CSTB2022TIAD-KPX0101)Science and Technology Research and Development Program of China State Railway Group Co.,Ltd.(No.N2023G045)。
文摘The uplift resistance of the soil overlying shield tunnels significantly impacts their anti-floating stability.However,research on uplift resistance concerning special-shaped shield tunnels is limited.This study combines numerical simulation with machine learning techniques to explore this issue.It presents a summary of special-shaped tunnel geometries and introduces a shape coefficient.Through the finite element software,Plaxis3D,the study simulates six key parameters—shape coefficient,burial depth ratio,tunnel’s longest horizontal length,internal friction angle,cohesion,and soil submerged bulk density—that impact uplift resistance across different conditions.Employing XGBoost and ANN methods,the feature importance of each parameter was analyzed based on the numerical simulation results.The findings demonstrate that a tunnel shape more closely resembling a circle leads to reduced uplift resistance in the overlying soil,whereas other parameters exhibit the contrary effects.Furthermore,the study reveals a diminishing trend in the feature importance of buried depth ratio,internal friction angle,tunnel longest horizontal length,cohesion,soil submerged bulk density,and shape coefficient in influencing uplift resistance.
文摘Federated Learning(FL)has become a leading decentralized solution that enables multiple clients to train a model in a collaborative environment without directly sharing raw data,making it suitable for privacy-sensitive applications such as healthcare,finance,and smart systems.As the field continues to evolve,the research field has become more complex and scattered,covering different system designs,training methods,and privacy techniques.This survey is organized around the three core challenges:how the data is distributed,how models are synchronized,and how to defend against attacks.It provides a structured and up-to-date review of FL research from 2023 to 2025,offering a unified taxonomy that categorizes works by data distribution(Horizontal FL,Vertical FL,Federated Transfer Learning,and Personalized FL),training synchronization(synchronous and asynchronous FL),optimization strategies,and threat models(data leakage and poisoning attacks).In particular,we summarize the latest contributions in Vertical FL frameworks for secure multi-party learning,communication-efficient Horizontal FL,and domain-adaptive Federated Transfer Learning.Furthermore,we examine synchronization techniques addressing system heterogeneity,including straggler mitigation in synchronous FL and staleness management in asynchronous FL.The survey covers security threats in FL,such as gradient inversion,membership inference,and poisoning attacks,as well as their defense strategies that include privacy-preserving aggregation and anomaly detection.The paper concludes by outlining unresolved issues and highlighting challenges in handling personalized models,scalability,and real-world adoption.
基金supported by a grant(No.CRPG-25-2054)under the Cybersecurity Research and Innovation Pioneers Initiative,provided by the National Cybersecurity Authority(NCA)in the Kingdom of Saudi Arabia.
文摘Split Learning(SL)has been promoted as a promising collaborative machine learning technique designed to address data privacy and resource efficiency.Specifically,neural networks are divided into client and server subnetworks in order to mitigate the exposure of sensitive data and reduce the overhead on client devices,thereby making SL particularly suitable for resource-constrained devices.Although SL prevents the direct transmission of raw data,it does not alleviate entirely the risk of privacy breaches.In fact,the data intermediately transmitted to the server sub-model may include patterns or information that could reveal sensitive data.Moreover,achieving a balance between model utility and data privacy has emerged as a challenging problem.In this article,we propose a novel defense approach that combines:(i)Adversarial learning,and(ii)Network channel pruning.In particular,the proposed adversarial learning approach is specifically designed to reduce the risk of private data exposure while maintaining high performance for the utility task.On the other hand,the suggested channel pruning enables the model to adaptively adjust and reactivate pruned channels while conducting adversarial training.The integration of these two techniques reduces the informativeness of the intermediate data transmitted by the client sub-model,thereby enhancing its robustness against attribute inference attacks without adding significant computational overhead,making it wellsuited for IoT devices,mobile platforms,and Internet of Vehicles(IoV)scenarios.The proposed defense approach was evaluated using EfficientNet-B0,a widely adopted compact model,along with three benchmark datasets.The obtained results showcased its superior defense capability against attribute inference attacks compared to existing state-of-the-art methods.This research’s findings demonstrated the effectiveness of the proposed channel pruning-based adversarial training approach in achieving the intended compromise between utility and privacy within SL frameworks.In fact,the classification accuracy attained by the attackers witnessed a drastic decrease of 70%.
文摘Knowledge distillation has become a standard technique for compressing large language models into efficient student models,but existing methods often struggle to balance prediction accuracy with explanation quality.Recent approaches such as Distilling Step-by-Step(DSbS)introduce explanation supervision,yet they apply it in a uniform manner that may not fully exploit the different learning dynamics of prediction and explanation.In this work,we propose a task-structured curriculum learning(TSCL)framework that structures training into three sequential phases:(i)prediction-only,to establish stable feature representations;(ii)joint prediction-explanation,to align task outputs with rationale generation;and(iii)explanation-only,to refine the quality of rationales.This design provides a simple but effective modification to DSbS,requiring no architectural changes and adding negligible training cost.We justify the phase scheduling with ablation studies and convergence analysis,showing that an initial prediction-heavy stage followed by a balanced joint phase improves both stability and explanation alignment.Extensive experiments on five datasets(e-SNLI,ANLI,CommonsenseQA,SVAMP,and MedNLI)demonstrate that TSCL consistently outperforms strong baselines,achieving gains of+1.7-2.6 points in accuracy and 0.8-1.2 in ROUGE-L,corresponding to relative error reductions of up to 21%.Beyond lexical metrics,human evaluation and ERASERstyle faithfulness diagnostics confirm that TSCL produces more faithful and informative explanations.Comparative training curves further reveal faster convergence and lower variance across seeds.Efficiency analysis shows less than 3%overhead in wall-clock training time and no additional inference cost,making the approach practical for realworld deployment.This study demonstrates that a simple task-structured curriculum can significantly improve the effectiveness of knowledge distillation.By separating and sequencing objectives,TSCL achieves a better balance between accuracy,stability,and explanation quality.The framework generalizes across domains,including medical NLI,and offers a principled recipe for future applications in multimodal reasoning and reinforcement learning.
文摘Latest digital advancements have intensified the necessity for adaptive,data-driven and socially-centered learning ecosystems.This paper presents the formulation of a cross-platform,innovative,gamified and personalized Learning Ecosystem,which integrates 3D/VR environments,as well as machine learning algorithms,and business intelligence frameworks to enhance learner-centered education and inferenced decision-making.This Learning System makes use of immersive,analytically assessed virtual learning spaces,therefore facilitating real-time monitoring of not just learning performance,but also overall engagement and behavioral patterns,via a comprehensive set of sustainability-oriented ESG-aligned Key Performance Indicators(KPIs).Machine learning models support predictive analysis,personalized feedback,and hybrid recommendation mechanisms,whilst dedicated dashboards translate complex educational data into actionable insights for all Use Cases of the System(Educational Institutions,Educators and Learners).Additionally,the presented Learning System introduces a structured Mentoring and Consulting Subsystem,thence reinforcing human-centered guidance alongside automated intelligence.The Platform’s modular architecture and simulation-centered evaluation approach actively support personalized,and continuously optimized learning pathways.Thence,it exemplifies a mature,adaptive Learning Ecosystem,supporting immersive technologies,analytics,and pedagogical support,hence,contributing to contemporary digital learning innovation and sociotechnical transformation in education.
文摘Automated grading of dandruff severity is a clinically significant but challenging task due to the inherent ordinal nature of severity levels and the high prevalence of label noise from subjective expert annotations.Standard classification methods fail to address these dual challenges,limiting their real-world performance.In this paper,a novel,three-phase training framework is proposed that learns a robust ordinal classifier directly from noisy labels.The approach synergistically combines a rank-based ordinal regression backbone with a cooperative,semi-supervised learning strategy to dynamically partition the data into clean and noisy subsets.A hybrid training objective is then employed,applying a supervised ordinal loss to the clean set.The noisy set is simultaneously trained using a dualobjective that combines a semi-supervised ordinal loss with a parallel,label-agnostic contrastive loss.This design allows themodel to learn fromthe entire noisy subset while using contrastive learning to mitigate the risk of error propagation frompotentially corrupt supervision.Extensive experiments on a new,large-scale,multi-site clinical dataset validate our approach.Themethod achieves state-of-the-art performance with 80.71%accuracy and a 76.86%F1-score,significantly outperforming existing approaches,including a 2.26%improvement over the strongest baseline method.This work provides not only a robust solution for a practical medical imaging problem but also a generalizable framework for other tasks plagued by noisy ordinal labels.
