This study proposes the use of the MERISE conceptual data model to create indicators for monitoring and evaluating the effectiveness of vocational training in the Republic of Congo. The importance of MERISE for struct...This study proposes the use of the MERISE conceptual data model to create indicators for monitoring and evaluating the effectiveness of vocational training in the Republic of Congo. The importance of MERISE for structuring and analyzing data is underlined, as it enables the measurement of the adequacy between training and the needs of the labor market. The innovation of the study lies in the adaptation of the MERISE model to the local context, the development of innovative indicators, and the integration of a participatory approach including all relevant stakeholders. Contextual adaptation and local innovation: The study suggests adapting MERISE to the specific context of the Republic of Congo, considering the local particularities of the labor market. Development of innovative indicators and new measurement tools: It proposes creating indicators to assess skills matching and employer satisfaction, which are crucial for evaluating the effectiveness of vocational training. Participatory approach and inclusion of stakeholders: The study emphasizes actively involving training centers, employers, and recruitment agencies in the evaluation process. This participatory approach ensures that the perspectives of all stakeholders are considered, leading to more relevant and practical outcomes. Using the MERISE model allows for: • Rigorous data structuring, organization, and standardization: Clearly defining entities and relationships facilitates data organization and standardization, crucial for effective data analysis. • Facilitation of monitoring, analysis, and relevant indicators: Developing both quantitative and qualitative indicators helps measure the effectiveness of training in relation to the labor market, allowing for a comprehensive evaluation. • Improved communication and common language: By providing a common language for different stakeholders, MERISE enhances communication and collaboration, ensuring that all parties have a shared understanding. The study’s approach and contribution to existing research lie in: • Structured theoretical and practical framework and holistic approach: The study offers a structured framework for data collection and analysis, covering both quantitative and qualitative aspects, thus providing a comprehensive view of the training system. • Reproducible methodology and international comparison: The proposed methodology can be replicated in other contexts, facilitating international comparison and the adoption of best practices. • Extension of knowledge and new perspective: By integrating a participatory approach and developing indicators adapted to local needs, the study extends existing research and offers new perspectives on vocational training evaluation.展开更多
In the realm of subsurface flow simulations,deep-learning-based surrogate models have emerged as a promising alternative to traditional simulation methods,especially in addressing complex optimization problems.However...In the realm of subsurface flow simulations,deep-learning-based surrogate models have emerged as a promising alternative to traditional simulation methods,especially in addressing complex optimization problems.However,a significant challenge lies in the necessity of numerous high-fidelity training simulations to construct these deep-learning models,which limits their application to field-scale problems.To overcome this limitation,we introduce a training procedure that leverages transfer learning with multi-fidelity training data to construct surrogate models efficiently.The procedure begins with the pre-training of the surrogate model using a relatively larger amount of data that can be efficiently generated from upscaled coarse-scale models.Subsequently,the model parameters are finetuned with a much smaller set of high-fidelity simulation data.For the cases considered in this study,this method leads to about a 75%reduction in total computational cost,in comparison with the traditional training approach,without any sacrifice of prediction accuracy.In addition,a dedicated well-control embedding model is introduced to the traditional U-Net architecture to improve the surrogate model's prediction accuracy,which is shown to be particularly effective when dealing with large-scale reservoir models under time-varying well control parameters.Comprehensive results and analyses are presented for the prediction of well rates,pressure and saturation states of a 3D synthetic reservoir system.Finally,the proposed procedure is applied to a field-scale production optimization problem.The trained surrogate model is shown to provide excellent generalization capabilities during the optimization process,in which the final optimized net-present-value is much higher than those from the training data ranges.展开更多
Background The Registrar Clinical Encounters in Training(ReCEnT)project is an Australian general practice vocational training programme with integrated and interdependent education and research functions.Trainees(regi...Background The Registrar Clinical Encounters in Training(ReCEnT)project is an Australian general practice vocational training programme with integrated and interdependent education and research functions.Trainees(registrars)contemporaneously document in-consultation clinical experience and actions.Objectives Using a realist lens,we elucidate the mechanisms underpinning project outcomes to answer questions around programme effectiveness,impacts,sustainability and the lessons and findings that are translatable to other primary care training programmes.Methods The context,input,process and product framework was used.As a means to understand the interactions between each of the interdependent components,it allows for inferences regarding causal mechanisms for specific outcomes.Results Context:ReCEnT occurs within an apprenticeship-like model of general practice vocational training entailing a central supervisor/apprentice relationship.ReCEnT has demystified the content and characteristics of registrar consultations.Input:multiple stakeholder involvement is both advantageous and a logistical challenge,with the programme’s success dependent on registrars,practices and training providers providing detailed and accurate data,with prompt subsequent processing.Process:contemporaneous consultation data collection in different stages of training constitutes a component of registrars’programmatic assessment.Product:individualised feedback provides educational benefit through reflection.Clinical and educational research questions can be addressed with resulting research translation feeding back into the programme model and government policy.Clinical behaviour change is also evaluated.Conclusion ReCEnT is unique,globally,in its scope and longevity(2010–present).Creation of meaningful,individualised feedback facilitates reflection and provides both immediate educational benefits and the substrate for further research,programme and policy design and targeted formal teaching and learning.展开更多
Social media data created a paradigm shift in assessing situational awareness during a natural disaster or emergencies such as wildfire, hurricane, tropical storm etc. Twitter as an emerging data source is an effectiv...Social media data created a paradigm shift in assessing situational awareness during a natural disaster or emergencies such as wildfire, hurricane, tropical storm etc. Twitter as an emerging data source is an effective and innovative digital platform to observe trend from social media users’ perspective who are direct or indirect witnesses of the calamitous event. This paper aims to collect and analyze twitter data related to the recent wildfire in California to perform a trend analysis by classifying firsthand and credible information from Twitter users. This work investigates tweets on the recent wildfire in California and classifies them based on witnesses into two types: 1) direct witnesses and 2) indirect witnesses. The collected and analyzed information can be useful for law enforcement agencies and humanitarian organizations for communication and verification of the situational awareness during wildfire hazards. Trend analysis is an aggregated approach that includes sentimental analysis and topic modeling performed through domain-expert manual annotation and machine learning. Trend analysis ultimately builds a fine-grained analysis to assess evacuation routes and provide valuable information to the firsthand emergency responders<span style="font-family:Verdana;">.</span>展开更多
This paper presents an procedure for purifying training data sets （i.e., past occurrences of slope failures） for inverse estimation on unobserved trigger factors of ＂different types of simultaneous slope failures＂...This paper presents an procedure for purifying training data sets （i.e., past occurrences of slope failures） for inverse estimation on unobserved trigger factors of ＂different types of simultaneous slope failures＂. Due to difficulties in pixel-by-pixel observations of trigger factors, as one of the measures, the authors had proposed an inverse analysis algorithm on trigger factors based on SEM （structural equation modeling）. Through a measurement equation, the trigger factor is inversely estimated, and a TFI （trigger factor influence） map can be also produced. As a subsequence subject, a purification procedure of training data set should be constructed to improve the accuracy of TFI map which depends on the representativeness of given training data sets of different types of slope failures. The proposed procedure resamples the matched pixels between original groups of past slope failures （i.e., surface slope failures, deep-seated slope failures, landslides） and classified three groups by K-means clustering for all pixels corresponding to those slope failures. For all cases of three types of slope failures, the improvement of success rates with respect to resampled training data sets was confirmed. As a final outcome, the differences between TFI maps produced by using original and resampled training data sets, respectively, are delineated on a DIF map （difference map） which is useful for analyzing trigger factor influence in terms of ＂risky- and safe-side assessment＂ sub-areas with respect to ＂different types of simultaneous slope failures＂.展开更多
When designing large-sized complex machinery products, the design focus is always on the overall per- formance; however, there exist no design theory and method based on performance driven. In view of the defi- ciency...When designing large-sized complex machinery products, the design focus is always on the overall per- formance; however, there exist no design theory and method based on performance driven. In view of the defi- ciency of the existing design theory, according to the performance features of complex mechanical products, the performance indices are introduced into the traditional design theory of ＂Requirement-Function-Structure＂ to construct a new five-domain design theory of ＂Client Requirement-Function-Performance-Structure-Design Parameter＂. To support design practice based on this new theory, a product data model is established by using per- formance indices and the mapping relationship between them and the other four domains. When the product data model is applied to high-speed train design and combining the existing research result and relevant standards, the corresponding data model and its structure involving five domains of high-speed trains are established, which can provide technical support for studying the relationships between typical performance indices and design parame- ters and the fast achievement of a high-speed train scheme design. The five domains provide a reference for the design specification and evaluation criteria of high speed train and a new idea for the train＇s parameter design.展开更多
Sand dust belts span approximately one-fifth of the global land surface.In these regions,dust tends to settle on vegetation surfaces,altering the observed reflectance and affecting remote sensing detections.To enhance...Sand dust belts span approximately one-fifth of the global land surface.In these regions,dust tends to settle on vegetation surfaces,altering the observed reflectance and affecting remote sensing detections.To enhance the accuracy of maize growth monitoring in dust-affected regions,this study aims to quantify the effect of sand dust retention on maize during the tasseling stage in the Kashgar Prefecture,Xinjiang Uygur Autonomous Region,China,by analyzing changes in canopy reflectance and vegetation indices.First,field sampling was conducted to measure the key canopy structure parameters and dust retention levels of maize,and laboratory spectral measurements were performed on leaf spectral properties under gradient dust retention.The measured data were then used to drive the LargE-Scale remote sensing data and image Simulation framework(LESS)model for simulating realistic maize canopy spectra across different dust levels,with validation against Sentinel-2 imagery.Second,on the basis of the simulated and satellite-derived spectra,the dust resistance of 36 common vegetation indices was systematically evaluated,and new robust dust-resistant indices were developed.The results showed that compared with dust-free maize,the canopy reflectance of dust-retained maize followed an increase–decrease–increase pattern,with critical turning points at 735 and 1325 nm.The maximum reflectance difference of–0.11755(change rate:29.002%)occurred within the 735–1325 nm range at 24 g/m^(2)dust retention,and the minimum reflectance difference of 0.04285(change rate:148.950%)was observed in the 350–735 nm range under the same dust retention level.Among the 36 vegetation indices,only the global environment monitoring index(GEMI)and the ratio of transformed chlorophyll absorption in reflectance index to optimized soil-adjusted vegetation index(TCARI/OSAVI)exhibited dust resistance,with GEMI being effective below 6 g/m^(2)and TCARI/OSAVI remaining stable across all levels(average ratio:0.970).The newly developed indices in this study,(RE3–RE2)/(NIR–RE2),(RE3–RE2)/(RE4–RE2),and(NIR–RE2)/(RE4–RE2),retained values within the predefined dust-resistant range over the full dust retention levels of 0–24 g/m^(2),thus showing a more stable dust resistance compared with the commonly used 36 vegetation indices.Specially,(RE3–RE2)/(RE4–RE2)performed the most robustly in Sentinel-2 imagery,that is,58.020%of pixels were within the dust-resistant range,and an average ratio of 0.937 was obtained for the original-spectra index.This study provides a scientific basis for crop monitoring and management in dust-affected regions.展开更多
In recent years,deep learning-based signal recognition technology has gained attention and emerged as an important approach for safeguarding the electromagnetic environment.However,training deep learning-based classif...In recent years,deep learning-based signal recognition technology has gained attention and emerged as an important approach for safeguarding the electromagnetic environment.However,training deep learning-based classifiers on large signal datasets with redundant samples requires significant memory and high costs.This paper proposes a support databased core-set selection method(SD)for signal recognition,aiming to screen a representative subset that approximates the large signal dataset.Specifically,this subset can be identified by employing the labeled information during the early stages of model training,as some training samples are labeled as supporting data frequently.This support data is crucial for model training and can be found using a border sample selector.Simulation results demonstrate that the SD method minimizes the impact on model recognition performance while reducing the dataset size,and outperforms five other state-of-the-art core-set selection methods when the fraction of training sample kept is less than or equal to 0.3 on the RML2016.04C dataset or 0.5 on the RML22 dataset.The SD method is particularly helpful for signal recognition tasks with limited memory and computing resources.展开更多
The focus of this study is to explore the statis-tical distribution models of high-speed railway (HSR) train delays. Based on actual HSR operational data, the delay causes and their classification, delay frequency, nu...The focus of this study is to explore the statis-tical distribution models of high-speed railway (HSR) train delays. Based on actual HSR operational data, the delay causes and their classification, delay frequency, number of affected trains, and space–time delay distributions are discussed. Eleven types of delay events are classified, and a detailed analysis of delay distribution for each classifica-tion is presented. Models of delay probability delay prob-ability distribution for each cause are proposed. Different distribution functions, including the lognormal, exponen-tial, gamma, uniform, logistic, and normal distribution, were selected to estimate and model delay patterns. The most appropriate distribution, which can approximate the delay duration corresponding to each cause, is derived. Subsequently, the Kolmogorov–Smirnov (K–S) test was used to test the goodness of fit of different train delay distribution models and the associated parameter values. The test results show that the distribution of the test data is consistent with that of the selected models. The fitting distribution models show the execution effect of the timetable and help in finding out the potential conflicts in real-time train operations.展开更多
Large Language Models (LLMs) have revolutionized Generative Artificial Intelligence (GenAI) tasks, becoming an integral part of various applications in society, including text generation, translation, summarization, a...Large Language Models (LLMs) have revolutionized Generative Artificial Intelligence (GenAI) tasks, becoming an integral part of various applications in society, including text generation, translation, summarization, and more. However, their widespread usage emphasizes the critical need to enhance their security posture to ensure the integrity and reliability of their outputs and minimize harmful effects. Prompt injections and training data poisoning attacks are two of the most prominent vulnerabilities in LLMs, which could potentially lead to unpredictable and undesirable behaviors, such as biased outputs, misinformation propagation, and even malicious content generation. The Common Vulnerability Scoring System (CVSS) framework provides a standardized approach to capturing the principal characteristics of vulnerabilities, facilitating a deeper understanding of their severity within the security and AI communities. By extending the current CVSS framework, we generate scores for these vulnerabilities such that organizations can prioritize mitigation efforts, allocate resources effectively, and implement targeted security measures to defend against potential risks.展开更多
文摘This study proposes the use of the MERISE conceptual data model to create indicators for monitoring and evaluating the effectiveness of vocational training in the Republic of Congo. The importance of MERISE for structuring and analyzing data is underlined, as it enables the measurement of the adequacy between training and the needs of the labor market. The innovation of the study lies in the adaptation of the MERISE model to the local context, the development of innovative indicators, and the integration of a participatory approach including all relevant stakeholders. Contextual adaptation and local innovation: The study suggests adapting MERISE to the specific context of the Republic of Congo, considering the local particularities of the labor market. Development of innovative indicators and new measurement tools: It proposes creating indicators to assess skills matching and employer satisfaction, which are crucial for evaluating the effectiveness of vocational training. Participatory approach and inclusion of stakeholders: The study emphasizes actively involving training centers, employers, and recruitment agencies in the evaluation process. This participatory approach ensures that the perspectives of all stakeholders are considered, leading to more relevant and practical outcomes. Using the MERISE model allows for: • Rigorous data structuring, organization, and standardization: Clearly defining entities and relationships facilitates data organization and standardization, crucial for effective data analysis. • Facilitation of monitoring, analysis, and relevant indicators: Developing both quantitative and qualitative indicators helps measure the effectiveness of training in relation to the labor market, allowing for a comprehensive evaluation. • Improved communication and common language: By providing a common language for different stakeholders, MERISE enhances communication and collaboration, ensuring that all parties have a shared understanding. The study’s approach and contribution to existing research lie in: • Structured theoretical and practical framework and holistic approach: The study offers a structured framework for data collection and analysis, covering both quantitative and qualitative aspects, thus providing a comprehensive view of the training system. • Reproducible methodology and international comparison: The proposed methodology can be replicated in other contexts, facilitating international comparison and the adoption of best practices. • Extension of knowledge and new perspective: By integrating a participatory approach and developing indicators adapted to local needs, the study extends existing research and offers new perspectives on vocational training evaluation.
基金funding support from the National Natural Science Foundation of China(No.52204065,No.ZX20230398)supported by a grant from the Human Resources Development Program(No.20216110100070)of the Korea Institute of Energy Technology Evaluation and Planning(KETEP)。
文摘In the realm of subsurface flow simulations,deep-learning-based surrogate models have emerged as a promising alternative to traditional simulation methods,especially in addressing complex optimization problems.However,a significant challenge lies in the necessity of numerous high-fidelity training simulations to construct these deep-learning models,which limits their application to field-scale problems.To overcome this limitation,we introduce a training procedure that leverages transfer learning with multi-fidelity training data to construct surrogate models efficiently.The procedure begins with the pre-training of the surrogate model using a relatively larger amount of data that can be efficiently generated from upscaled coarse-scale models.Subsequently,the model parameters are finetuned with a much smaller set of high-fidelity simulation data.For the cases considered in this study,this method leads to about a 75%reduction in total computational cost,in comparison with the traditional training approach,without any sacrifice of prediction accuracy.In addition,a dedicated well-control embedding model is introduced to the traditional U-Net architecture to improve the surrogate model's prediction accuracy,which is shown to be particularly effective when dealing with large-scale reservoir models under time-varying well control parameters.Comprehensive results and analyses are presented for the prediction of well rates,pressure and saturation states of a 3D synthetic reservoir system.Finally,the proposed procedure is applied to a field-scale production optimization problem.The trained surrogate model is shown to provide excellent generalization capabilities during the optimization process,in which the final optimized net-present-value is much higher than those from the training data ranges.
文摘Background The Registrar Clinical Encounters in Training(ReCEnT)project is an Australian general practice vocational training programme with integrated and interdependent education and research functions.Trainees(registrars)contemporaneously document in-consultation clinical experience and actions.Objectives Using a realist lens,we elucidate the mechanisms underpinning project outcomes to answer questions around programme effectiveness,impacts,sustainability and the lessons and findings that are translatable to other primary care training programmes.Methods The context,input,process and product framework was used.As a means to understand the interactions between each of the interdependent components,it allows for inferences regarding causal mechanisms for specific outcomes.Results Context:ReCEnT occurs within an apprenticeship-like model of general practice vocational training entailing a central supervisor/apprentice relationship.ReCEnT has demystified the content and characteristics of registrar consultations.Input:multiple stakeholder involvement is both advantageous and a logistical challenge,with the programme’s success dependent on registrars,practices and training providers providing detailed and accurate data,with prompt subsequent processing.Process:contemporaneous consultation data collection in different stages of training constitutes a component of registrars’programmatic assessment.Product:individualised feedback provides educational benefit through reflection.Clinical and educational research questions can be addressed with resulting research translation feeding back into the programme model and government policy.Clinical behaviour change is also evaluated.Conclusion ReCEnT is unique,globally,in its scope and longevity(2010–present).Creation of meaningful,individualised feedback facilitates reflection and provides both immediate educational benefits and the substrate for further research,programme and policy design and targeted formal teaching and learning.
文摘Social media data created a paradigm shift in assessing situational awareness during a natural disaster or emergencies such as wildfire, hurricane, tropical storm etc. Twitter as an emerging data source is an effective and innovative digital platform to observe trend from social media users’ perspective who are direct or indirect witnesses of the calamitous event. This paper aims to collect and analyze twitter data related to the recent wildfire in California to perform a trend analysis by classifying firsthand and credible information from Twitter users. This work investigates tweets on the recent wildfire in California and classifies them based on witnesses into two types: 1) direct witnesses and 2) indirect witnesses. The collected and analyzed information can be useful for law enforcement agencies and humanitarian organizations for communication and verification of the situational awareness during wildfire hazards. Trend analysis is an aggregated approach that includes sentimental analysis and topic modeling performed through domain-expert manual annotation and machine learning. Trend analysis ultimately builds a fine-grained analysis to assess evacuation routes and provide valuable information to the firsthand emergency responders<span style="font-family:Verdana;">.</span>
文摘This paper presents an procedure for purifying training data sets （i.e., past occurrences of slope failures） for inverse estimation on unobserved trigger factors of ＂different types of simultaneous slope failures＂. Due to difficulties in pixel-by-pixel observations of trigger factors, as one of the measures, the authors had proposed an inverse analysis algorithm on trigger factors based on SEM （structural equation modeling）. Through a measurement equation, the trigger factor is inversely estimated, and a TFI （trigger factor influence） map can be also produced. As a subsequence subject, a purification procedure of training data set should be constructed to improve the accuracy of TFI map which depends on the representativeness of given training data sets of different types of slope failures. The proposed procedure resamples the matched pixels between original groups of past slope failures （i.e., surface slope failures, deep-seated slope failures, landslides） and classified three groups by K-means clustering for all pixels corresponding to those slope failures. For all cases of three types of slope failures, the improvement of success rates with respect to resampled training data sets was confirmed. As a final outcome, the differences between TFI maps produced by using original and resampled training data sets, respectively, are delineated on a DIF map （difference map） which is useful for analyzing trigger factor influence in terms of ＂risky- and safe-side assessment＂ sub-areas with respect to ＂different types of simultaneous slope failures＂.
基金Supported by National Natural Science Foundation of China(Grant Nos.51275432,51505390)Sichuan Application Foundation Projects(Grant No.2016JY0098)Independent Research Project of TPL(Grant No.TPL1501)
文摘When designing large-sized complex machinery products, the design focus is always on the overall per- formance; however, there exist no design theory and method based on performance driven. In view of the defi- ciency of the existing design theory, according to the performance features of complex mechanical products, the performance indices are introduced into the traditional design theory of ＂Requirement-Function-Structure＂ to construct a new five-domain design theory of ＂Client Requirement-Function-Performance-Structure-Design Parameter＂. To support design practice based on this new theory, a product data model is established by using per- formance indices and the mapping relationship between them and the other four domains. When the product data model is applied to high-speed train design and combining the existing research result and relevant standards, the corresponding data model and its structure involving five domains of high-speed trains are established, which can provide technical support for studying the relationships between typical performance indices and design parame- ters and the fast achievement of a high-speed train scheme design. The five domains provide a reference for the design specification and evaluation criteria of high speed train and a new idea for the train＇s parameter design.
基金supported by the Fundamental Research Funds for the Central Universities(N2001020)the National Natural Science Foundation of China(41201359).
文摘Sand dust belts span approximately one-fifth of the global land surface.In these regions,dust tends to settle on vegetation surfaces,altering the observed reflectance and affecting remote sensing detections.To enhance the accuracy of maize growth monitoring in dust-affected regions,this study aims to quantify the effect of sand dust retention on maize during the tasseling stage in the Kashgar Prefecture,Xinjiang Uygur Autonomous Region,China,by analyzing changes in canopy reflectance and vegetation indices.First,field sampling was conducted to measure the key canopy structure parameters and dust retention levels of maize,and laboratory spectral measurements were performed on leaf spectral properties under gradient dust retention.The measured data were then used to drive the LargE-Scale remote sensing data and image Simulation framework(LESS)model for simulating realistic maize canopy spectra across different dust levels,with validation against Sentinel-2 imagery.Second,on the basis of the simulated and satellite-derived spectra,the dust resistance of 36 common vegetation indices was systematically evaluated,and new robust dust-resistant indices were developed.The results showed that compared with dust-free maize,the canopy reflectance of dust-retained maize followed an increase–decrease–increase pattern,with critical turning points at 735 and 1325 nm.The maximum reflectance difference of–0.11755(change rate:29.002%)occurred within the 735–1325 nm range at 24 g/m^(2)dust retention,and the minimum reflectance difference of 0.04285(change rate:148.950%)was observed in the 350–735 nm range under the same dust retention level.Among the 36 vegetation indices,only the global environment monitoring index(GEMI)and the ratio of transformed chlorophyll absorption in reflectance index to optimized soil-adjusted vegetation index(TCARI/OSAVI)exhibited dust resistance,with GEMI being effective below 6 g/m^(2)and TCARI/OSAVI remaining stable across all levels(average ratio:0.970).The newly developed indices in this study,(RE3–RE2)/(NIR–RE2),(RE3–RE2)/(RE4–RE2),and(NIR–RE2)/(RE4–RE2),retained values within the predefined dust-resistant range over the full dust retention levels of 0–24 g/m^(2),thus showing a more stable dust resistance compared with the commonly used 36 vegetation indices.Specially,(RE3–RE2)/(RE4–RE2)performed the most robustly in Sentinel-2 imagery,that is,58.020%of pixels were within the dust-resistant range,and an average ratio of 0.937 was obtained for the original-spectra index.This study provides a scientific basis for crop monitoring and management in dust-affected regions.
基金supported by National Natural Science Foundation of China(62371098)Natural Science Foundation of Sichuan Province(2023NSFSC1422)+1 种基金National Key Research and Development Program of China(2021YFB2900404)Central Universities of South west Minzu University(ZYN2022032).
文摘In recent years,deep learning-based signal recognition technology has gained attention and emerged as an important approach for safeguarding the electromagnetic environment.However,training deep learning-based classifiers on large signal datasets with redundant samples requires significant memory and high costs.This paper proposes a support databased core-set selection method(SD)for signal recognition,aiming to screen a representative subset that approximates the large signal dataset.Specifically,this subset can be identified by employing the labeled information during the early stages of model training,as some training samples are labeled as supporting data frequently.This support data is crucial for model training and can be found using a border sample selector.Simulation results demonstrate that the SD method minimizes the impact on model recognition performance while reducing the dataset size,and outperforms five other state-of-the-art core-set selection methods when the fraction of training sample kept is less than or equal to 0.3 on the RML2016.04C dataset or 0.5 on the RML22 dataset.The SD method is particularly helpful for signal recognition tasks with limited memory and computing resources.
基金supported by the National Key R&D Plan (No.2017YFB1200701)National Nature Science Foundation of China (No.U1834209 and 71871188)the support of the Railways Technology Development Plan of China Railway Corporation (No.2016X008-J)supported by State Key Lab of Railway Control and Safety Open Topics Fund (No.RCS2019K007)
文摘The focus of this study is to explore the statis-tical distribution models of high-speed railway (HSR) train delays. Based on actual HSR operational data, the delay causes and their classification, delay frequency, number of affected trains, and space–time delay distributions are discussed. Eleven types of delay events are classified, and a detailed analysis of delay distribution for each classifica-tion is presented. Models of delay probability delay prob-ability distribution for each cause are proposed. Different distribution functions, including the lognormal, exponen-tial, gamma, uniform, logistic, and normal distribution, were selected to estimate and model delay patterns. The most appropriate distribution, which can approximate the delay duration corresponding to each cause, is derived. Subsequently, the Kolmogorov–Smirnov (K–S) test was used to test the goodness of fit of different train delay distribution models and the associated parameter values. The test results show that the distribution of the test data is consistent with that of the selected models. The fitting distribution models show the execution effect of the timetable and help in finding out the potential conflicts in real-time train operations.
文摘Large Language Models (LLMs) have revolutionized Generative Artificial Intelligence (GenAI) tasks, becoming an integral part of various applications in society, including text generation, translation, summarization, and more. However, their widespread usage emphasizes the critical need to enhance their security posture to ensure the integrity and reliability of their outputs and minimize harmful effects. Prompt injections and training data poisoning attacks are two of the most prominent vulnerabilities in LLMs, which could potentially lead to unpredictable and undesirable behaviors, such as biased outputs, misinformation propagation, and even malicious content generation. The Common Vulnerability Scoring System (CVSS) framework provides a standardized approach to capturing the principal characteristics of vulnerabilities, facilitating a deeper understanding of their severity within the security and AI communities. By extending the current CVSS framework, we generate scores for these vulnerabilities such that organizations can prioritize mitigation efforts, allocate resources effectively, and implement targeted security measures to defend against potential risks.
