Radio frequency quadrupoles(RFQs),which are crucial components of proton injectors,significantly affect the performance of proton accelerator facilities.An RFQ with a high frequency of 714 MHz dedicated to compact pro...Radio frequency quadrupoles(RFQs),which are crucial components of proton injectors,significantly affect the performance of proton accelerator facilities.An RFQ with a high frequency of 714 MHz dedicated to compact proton injectors for medi-cal applications is designed in this study.The RFQ is designed to accelerate proton beams from 50 keV to 4 MeV within a short length of 2 m and can be matched closely with the downstream drift tube linac to capture more particles through a preliminary optimization.To develop an advanced RFQ,challenging techniques,including fabrication and tuning method,must be evaluated and verified using a prototype.An aluminium prototype is derived from the conceptual design of the RFQ and then redesigned to confirm the radio frequency performance,fabrication procedure,and feasibility of the tuning algorithm.Eventually,a new tuning algorithm based on the response matrix and least-squares method is developed,which yields favorable results based on the prototype,i.e.,the errors of the dipole and quadrupole components reduced to a low level after several tuning iterations.Benefiting from the conceptual design and techniques obtained from the prototype,the formal mechanical design of the 2-m RFQ is ready for the next manufacturing step.展开更多
We report a passive mode-locked fiber laser that can realize single-wavelength tuning and multi-wavelength spacing tuning simultaneously.The tuning range is from 1528 nm–1560 nm,and up to three bands of soliton state...We report a passive mode-locked fiber laser that can realize single-wavelength tuning and multi-wavelength spacing tuning simultaneously.The tuning range is from 1528 nm–1560 nm,and up to three bands of soliton states can be output at the same time.These results are confirmed by a nonlinear Schrodinger equation model based on the split-step Fourier method.In addition,we reveal a way to transform the multi-wavelength soliton state into the Q-switched mode-locked state,which is period doubling.These results will promote the development of optical communication,optical sensing and multi-signal pulse emission.展开更多
Based on the numerical solution of the time-dependent Dirac equation,we propose a method to observe in real time the ac Stark shift of a vacuum driven by an ultra-intense laser field.By overlapping the ultra-intense p...Based on the numerical solution of the time-dependent Dirac equation,we propose a method to observe in real time the ac Stark shift of a vacuum driven by an ultra-intense laser field.By overlapping the ultra-intense pump pulse with another zeptosecond probe pulse whose photon energy is smaller than 2mc^(2),electron–positron pair creation can be controlled by tuning the time delay between the pump and probe pulses.Since the pair creation rate depends sensitively on the instantaneous vacuum potential,one can reconstruct the ac Stark shift of the vacuum potential according to the time-delay-dependent pair creation rate.展开更多
With the rapid development of large megawatt wind turbines,the operation environment of wind turbine towers(WTTs)has become increasingly complex.In particular,seismic excitation can create a resonance response and cau...With the rapid development of large megawatt wind turbines,the operation environment of wind turbine towers(WTTs)has become increasingly complex.In particular,seismic excitation can create a resonance response and cause excessive vibration of the WTT.To investigate the vibration attenuation performance of the WTT under seismic excitations,a novel passive vibration control device,called a prestressed tuned mass damper(PS-TMD),is presented in this study.First,a mathematical model is established based on structural dynamics under seismic excitation.Then,the mathematical analytical expression of the dynamic coefficient is deduced,and the parameter design method is obtained by system tuning optimization.Next,based on a theoretical analysis and parameter design,the numerical results showed that the PS-TMD was able to effectively mitigate the resonance under the harmonic basal acceleration.Finally,the time-history analysis method is used to verify the effectiveness of the traditional pendulum tuned mass damper(PTMD)and the novel PS-TMD device,and the results indicate that the vibration attenuation performance of the PS-TMD is better than the PTMD.In addition,the PS-TMD avoids the nonlinear effect due to the large oscillation angle,and has the potential to dissipate hysteretic energy under seismic excitation.展开更多
Glaucoma is a leading cause of irreve rsible blindness wo rldwide,and previous studies have shown that,in addition to affecting the eyes,it also causes abnormalities in the brain.However,it is not yet clear how the pr...Glaucoma is a leading cause of irreve rsible blindness wo rldwide,and previous studies have shown that,in addition to affecting the eyes,it also causes abnormalities in the brain.However,it is not yet clear how the primary visual cortex(V1)is altered in glaucoma.This study used DBA/2J mice as a model for spontaneous secondary glaucoma.The aim of the study was to compare the electrophysiological and histomorphological chara cteristics of neurons in the V1between 9-month-old DBA/2J mice and age-matched C57BL/6J mice.We conducted single-unit recordings in the V1 of light-anesthetized mice to measure the visually induced responses,including single-unit spiking and gamma band oscillations.The morphology of layerⅡ/Ⅲneurons was determined by neuronal nuclear antigen staining and Nissl staining of brain tissue sections.Eighty-seven neurons from eight DBA/2J mice and eighty-one neurons from eight C57BL/6J mice were examined.Compared with the C57BL/6J group,V1 neurons in the DBA/2J group exhibited weaker visual tuning and impaired spatial summation.Moreove r,fewer neuro ns were observed in the V1 of DBA/2J mice compared with C57BL/6J mice.These findings suggest that DBA/2J mice have fewer neurons in the VI compared with C57BL/6J mice,and that these neurons have impaired visual tuning.Our findings provide a better understanding of the pathological changes that occur in V1 neuron function and morphology in the DBA/2J mouse model.This study might offer some innovative perspectives regarding the treatment of glaucoma.展开更多
A highly sensitive light-induced thermoelectric spectroscopy(LITES)sensor based on a multi-pass cell(MPC)with dense spot pattern and a novel quartz tuning fork(QTF)with low resonance frequency is reported in this manu...A highly sensitive light-induced thermoelectric spectroscopy(LITES)sensor based on a multi-pass cell(MPC)with dense spot pattern and a novel quartz tuning fork(QTF)with low resonance frequency is reported in this manuscript.An erbi-um-doped fiber amplifier(EDFA)was employed to amplify the output optical power so that the signal level was further enhanced.The optical path length(OPL)and the ratio of optical path length to volume(RLV)of the MPC is 37.7 m and 13.8 cm^(-2),respectively.A commercial QTF and a self-designed trapezoidal-tip QTF with low frequency of 9461.83 Hz were used as the detectors of the sensor,respectively.The target gas selected to test the performance of the system was acetylene(C2H2).When the optical power was constant at 1000 mW,the minimum detection limit(MDL)of the C2H2-LITES sensor can be achieved 48.3 ppb when using the commercial QTF and 24.6 ppb when using the trapezoid-al-tip QTF.An improvement of the detection performance by a factor of 1.96 was achieved after replacing the commer-cial QTF with the trapezoidal-tip QTF.展开更多
Fraud of credit cards is a major issue for financial organizations and individuals.As fraudulent actions become more complex,a demand for better fraud detection systems is rising.Deep learning approaches have shown pr...Fraud of credit cards is a major issue for financial organizations and individuals.As fraudulent actions become more complex,a demand for better fraud detection systems is rising.Deep learning approaches have shown promise in several fields,including detecting credit card fraud.However,the efficacy of these models is heavily dependent on the careful selection of appropriate hyperparameters.This paper introduces models that integrate deep learning models with hyperparameter tuning techniques to learn the patterns and relationships within credit card transaction data,thereby improving fraud detection.Three deep learning models:AutoEncoder(AE),Convolution Neural Network(CNN),and Long Short-Term Memory(LSTM)are proposed to investigate how hyperparameter adjustment impacts the efficacy of deep learning models used to identify credit card fraud.The experiments conducted on a European credit card fraud dataset using different hyperparameters and three deep learning models demonstrate that the proposed models achieve a tradeoff between detection rate and precision,leading these models to be effective in accurately predicting credit card fraud.The results demonstrate that LSTM significantly outperformed AE and CNN in terms of accuracy(99.2%),detection rate(93.3%),and area under the curve(96.3%).These proposed models have surpassed those of existing studies and are expected to make a significant contribution to the field of credit card fraud detection.展开更多
In order to improve the seismic performance of adjacent buildings,two types of tuned inerter damper(TID)damping systems for adjacent buildings are proposed,which are composed of springs,inerter devices and dampers in ...In order to improve the seismic performance of adjacent buildings,two types of tuned inerter damper(TID)damping systems for adjacent buildings are proposed,which are composed of springs,inerter devices and dampers in serial or in parallel.The dynamic equations of TID adjacent building damping systems were derived,and the H2 norm criterion was used to optimize and adjust them,so that the system had the optimum damping performance under white noise random excitation.Taking TID frequency ratio and damping ratio as optimization parameters,the optimum analytical solutions of the displacement frequency response of the undamped structure under white noise excitation were obtained.The results showed that compared with the classic TMD,TID could obtain a better damping effect in the adjacent buildings.Comparing the TIDs composed of serial or parallel,it was found that the parallel TIDs had more significant advantages in controlling the peak displacement frequency response,while the H2 norm of the displacement frequency response of the damping system under the coupling of serial TID was smaller.Taking the adjacent building composed of two ten-story frame structures as an example,the displacement and energy collection time history analysis of the adjacent building coupled with the optimum design parameter TIDs were carried out.It was found that TID had a better damping effect in the full-time range compared with the classic TMD.This paper also studied the potential power of TID in adjacent buildings,which can be converted into available power resources during earthquakes.展开更多
Sentence classification is the process of categorizing a sentence based on the context of the sentence.Sentence categorization requires more semantic highlights than other tasks,such as dependence parsing,which requir...Sentence classification is the process of categorizing a sentence based on the context of the sentence.Sentence categorization requires more semantic highlights than other tasks,such as dependence parsing,which requires more syntactic elements.Most existing strategies focus on the general semantics of a conversation without involving the context of the sentence,recognizing the progress and comparing impacts.An ensemble pre-trained language model was taken up here to classify the conversation sentences from the conversation corpus.The conversational sentences are classified into four categories:information,question,directive,and commission.These classification label sequences are for analyzing the conversation progress and predicting the pecking order of the conversation.Ensemble of Bidirectional Encoder for Representation of Transformer(BERT),Robustly Optimized BERT pretraining Approach(RoBERTa),Generative Pre-Trained Transformer(GPT),DistilBERT and Generalized Autoregressive Pretraining for Language Understanding(XLNet)models are trained on conversation corpus with hyperparameters.Hyperparameter tuning approach is carried out for better performance on sentence classification.This Ensemble of Pre-trained Language Models with a Hyperparameter Tuning(EPLM-HT)system is trained on an annotated conversation dataset.The proposed approach outperformed compared to the base BERT,GPT,DistilBERT and XLNet transformer models.The proposed ensemble model with the fine-tuned parameters achieved an F1_score of 0.88.展开更多
Hyperparameter tuning is a key step in developing high-performing machine learning models, but searching large hyperparameter spaces requires extensive computation using standard sequential methods. This work analyzes...Hyperparameter tuning is a key step in developing high-performing machine learning models, but searching large hyperparameter spaces requires extensive computation using standard sequential methods. This work analyzes the performance gains from parallel versus sequential hyperparameter optimization. Using scikit-learn’s Randomized SearchCV, this project tuned a Random Forest classifier for fake news detection via randomized grid search. Setting n_jobs to -1 enabled full parallelization across CPU cores. Results show the parallel implementation achieved over 5× faster CPU times and 3× faster total run times compared to sequential tuning. However, test accuracy slightly dropped from 99.26% sequentially to 99.15% with parallelism, indicating a trade-off between evaluation efficiency and model performance. Still, the significant computational gains allow more extensive hyperparameter exploration within reasonable timeframes, outweighing the small accuracy decrease. Further analysis could better quantify this trade-off across different models, tuning techniques, tasks, and hardware.展开更多
超高层建筑是现代城市建设的重要标志之一,其高度已经超过了传统建筑的极限。然而,随着建筑高度不断增加,地震的破坏力也越来越强,超高层建筑面临着更加严峻的安全挑战。因此,研究超高层建筑防震支撑系统技术非常重要,Tuned Mass Damper...超高层建筑是现代城市建设的重要标志之一,其高度已经超过了传统建筑的极限。然而,随着建筑高度不断增加,地震的破坏力也越来越强,超高层建筑面临着更加严峻的安全挑战。因此,研究超高层建筑防震支撑系统技术非常重要,Tuned Mass Damper(TMD)是一种被广泛研究和应用的超高层建筑防震支撑系统技术,TMD最初是在20世纪60年代提出的,最早应用于桥梁上,后来,TMD被引入到建筑领域,并得到广泛的应用。通过精确调节质量、阻尼和弹性等参数来削弱地震引起的建筑物减震效应,从而减少了建筑物因地震造成的损害和崩塌的风险.展开更多
Human Activity Recognition(HAR)has been made simple in recent years,thanks to recent advancements made in Artificial Intelligence(AI)techni-ques.These techniques are applied in several areas like security,surveillance,...Human Activity Recognition(HAR)has been made simple in recent years,thanks to recent advancements made in Artificial Intelligence(AI)techni-ques.These techniques are applied in several areas like security,surveillance,healthcare,human-robot interaction,and entertainment.Since wearable sensor-based HAR system includes in-built sensors,human activities can be categorized based on sensor values.Further,it can also be employed in other applications such as gait diagnosis,observation of children/adult’s cognitive nature,stroke-patient hospital direction,Epilepsy and Parkinson’s disease examination,etc.Recently-developed Artificial Intelligence(AI)techniques,especially Deep Learning(DL)models can be deployed to accomplish effective outcomes on HAR process.With this motivation,the current research paper focuses on designing Intelligent Hyperparameter Tuned Deep Learning-based HAR(IHPTDL-HAR)technique in healthcare environment.The proposed IHPTDL-HAR technique aims at recogniz-ing the human actions in healthcare environment and helps the patients in mana-ging their healthcare service.In addition,the presented model makes use of Hierarchical Clustering(HC)-based outlier detection technique to remove the out-liers.IHPTDL-HAR technique incorporates DL-based Deep Belief Network(DBN)model to recognize the activities of users.Moreover,Harris Hawks Opti-mization(HHO)algorithm is used for hyperparameter tuning of DBN model.Finally,a comprehensive experimental analysis was conducted upon benchmark dataset and the results were examined under different aspects.The experimental results demonstrate that the proposed IHPTDL-HAR technique is a superior per-former compared to other recent techniques under different measures.展开更多
Human pose estimation(HPE)is a procedure for determining the structure of the body pose and it is considered a challenging issue in the computer vision(CV)communities.HPE finds its applications in several fields namel...Human pose estimation(HPE)is a procedure for determining the structure of the body pose and it is considered a challenging issue in the computer vision(CV)communities.HPE finds its applications in several fields namely activity recognition and human-computer interface.Despite the benefits of HPE,it is still a challenging process due to the variations in visual appearances,lighting,occlusions,dimensionality,etc.To resolve these issues,this paper presents a squirrel search optimization with a deep convolutional neural network for HPE(SSDCNN-HPE)technique.The major intention of the SSDCNN-HPE technique is to identify the human pose accurately and efficiently.Primarily,the video frame conversion process is performed and pre-processing takes place via bilateral filtering-based noise removal process.Then,the EfficientNet model is applied to identify the body points of a person with no problem constraints.Besides,the hyperparameter tuning of the EfficientNet model takes place by the use of the squirrel search algorithm(SSA).In the final stage,the multiclass support vector machine(M-SVM)technique was utilized for the identification and classification of human poses.The design of bilateral filtering followed by SSA based EfficientNetmodel for HPE depicts the novelty of the work.To demonstrate the enhanced outcomes of the SSDCNN-HPE approach,a series of simulations are executed.The experimental results reported the betterment of the SSDCNN-HPE system over the recent existing techniques in terms of different measures.展开更多
Recently,computation offloading has become an effective method for overcoming the constraint of a mobile device(MD)using computationintensivemobile and offloading delay-sensitive application tasks to the remote cloud-...Recently,computation offloading has become an effective method for overcoming the constraint of a mobile device(MD)using computationintensivemobile and offloading delay-sensitive application tasks to the remote cloud-based data center.Smart city benefitted from offloading to edge point.Consider a mobile edge computing(MEC)network in multiple regions.They comprise N MDs and many access points,in which everyMDhasM independent real-time tasks.This study designs a new Task Offloading and Resource Allocation in IoT-based MEC using Deep Learning with Seagull Optimization(TORA-DLSGO)algorithm.The proposed TORA-DLSGO technique addresses the resource management issue in the MEC server,which enables an optimum offloading decision to minimize the system cost.In addition,an objective function is derived based on minimizing energy consumption subject to the latency requirements and restricted resources.The TORA-DLSGO technique uses the deep belief network(DBN)model for optimum offloading decision-making.Finally,the SGO algorithm is used for the parameter tuning of the DBN model.The simulation results exemplify that the TORA-DLSGO technique outperformed the existing model in reducing client overhead in the MEC systems with a maximum reward of 0.8967.展开更多
Remote sensing image(RSI)classifier roles a vital play in earth observation technology utilizing Remote sensing(RS)data are extremely exploited from both military and civil fields.More recently,as novel DL approaches ...Remote sensing image(RSI)classifier roles a vital play in earth observation technology utilizing Remote sensing(RS)data are extremely exploited from both military and civil fields.More recently,as novel DL approaches develop,techniques for RSI classifiers with DL have attained important breakthroughs,providing a new opportunity for the research and development of RSI classifiers.This study introduces an Improved Slime Mould Optimization with a graph convolutional network for the hyperspectral remote sensing image classification(ISMOGCN-HRSC)model.The ISMOGCN-HRSC model majorly concentrates on identifying and classifying distinct kinds of RSIs.In the presented ISMOGCN-HRSC model,the synergic deep learning(SDL)model is exploited to produce feature vectors.The GCN model is utilized for image classification purposes to identify the proper class labels of the RSIs.The ISMO algorithm is used to enhance the classification efficiency of the GCN method,which is derived by integrating chaotic concepts into the SMO algorithm.The experimental assessment of the ISMOGCN-HRSC method is tested using a benchmark dataset.展开更多
基金This work was supported by National Natural Science Foundation of China(No.12222513).
文摘Radio frequency quadrupoles(RFQs),which are crucial components of proton injectors,significantly affect the performance of proton accelerator facilities.An RFQ with a high frequency of 714 MHz dedicated to compact proton injectors for medi-cal applications is designed in this study.The RFQ is designed to accelerate proton beams from 50 keV to 4 MeV within a short length of 2 m and can be matched closely with the downstream drift tube linac to capture more particles through a preliminary optimization.To develop an advanced RFQ,challenging techniques,including fabrication and tuning method,must be evaluated and verified using a prototype.An aluminium prototype is derived from the conceptual design of the RFQ and then redesigned to confirm the radio frequency performance,fabrication procedure,and feasibility of the tuning algorithm.Eventually,a new tuning algorithm based on the response matrix and least-squares method is developed,which yields favorable results based on the prototype,i.e.,the errors of the dipole and quadrupole components reduced to a low level after several tuning iterations.Benefiting from the conceptual design and techniques obtained from the prototype,the formal mechanical design of the 2-m RFQ is ready for the next manufacturing step.
基金supported by the Zhejiang Provincial Natural Science Foundation of China(Grant No.LR20A050001)the National Natural Science Foundation of China(Grant Nos.12261131495 and 12275240)the Scientific Research and De-veloped Fund of Zhejiang A&F University(Grant No.2021FR0009).
文摘We report a passive mode-locked fiber laser that can realize single-wavelength tuning and multi-wavelength spacing tuning simultaneously.The tuning range is from 1528 nm–1560 nm,and up to three bands of soliton states can be output at the same time.These results are confirmed by a nonlinear Schrodinger equation model based on the split-step Fourier method.In addition,we reveal a way to transform the multi-wavelength soliton state into the Q-switched mode-locked state,which is period doubling.These results will promote the development of optical communication,optical sensing and multi-signal pulse emission.
基金supported by the National Natural Science Foundation of China(Grant Nos.12304341 and 11974419)the National Key R&D Program of China(Grant Nos.2021YFA1601700 and 2018YFA0404802)the Strategic Priority Research Program of Chinese Academy of Sciences(Grant No.XDA25051000).
文摘Based on the numerical solution of the time-dependent Dirac equation,we propose a method to observe in real time the ac Stark shift of a vacuum driven by an ultra-intense laser field.By overlapping the ultra-intense pump pulse with another zeptosecond probe pulse whose photon energy is smaller than 2mc^(2),electron–positron pair creation can be controlled by tuning the time delay between the pump and probe pulses.Since the pair creation rate depends sensitively on the instantaneous vacuum potential,one can reconstruct the ac Stark shift of the vacuum potential according to the time-delay-dependent pair creation rate.
基金Fundamental Research Funds for the National Natural Science Foundation of China under Grant No.52078084the Natural Science Foundation of Chongqing (cstc2021jcyj-msxmX0623)+2 种基金the 111 project of the Ministry of Educationthe Bureau of Foreign Experts of China under Grant No.B18062China Postdoctoral Science Foundation under Grant No.2021M690838。
文摘With the rapid development of large megawatt wind turbines,the operation environment of wind turbine towers(WTTs)has become increasingly complex.In particular,seismic excitation can create a resonance response and cause excessive vibration of the WTT.To investigate the vibration attenuation performance of the WTT under seismic excitations,a novel passive vibration control device,called a prestressed tuned mass damper(PS-TMD),is presented in this study.First,a mathematical model is established based on structural dynamics under seismic excitation.Then,the mathematical analytical expression of the dynamic coefficient is deduced,and the parameter design method is obtained by system tuning optimization.Next,based on a theoretical analysis and parameter design,the numerical results showed that the PS-TMD was able to effectively mitigate the resonance under the harmonic basal acceleration.Finally,the time-history analysis method is used to verify the effectiveness of the traditional pendulum tuned mass damper(PTMD)and the novel PS-TMD device,and the results indicate that the vibration attenuation performance of the PS-TMD is better than the PTMD.In addition,the PS-TMD avoids the nonlinear effect due to the large oscillation angle,and has the potential to dissipate hysteretic energy under seismic excitation.
基金supported by the STI 2030-Major Projects 2022ZD0208500(to DY)the National Natural Science Foundation of China,Nos.82072011(to YX),82121003(to DY),82271120(to YS)+2 种基金Sichuan Science and Technology Program,No.2022ZYD0066(to YS)a grant from Chinese Academy of Medical Science,No.2019-12M-5-032(to YS)the Fundamental Research Funds for the Central Universities,No.ZYGX2021YGLH219(to KC)。
文摘Glaucoma is a leading cause of irreve rsible blindness wo rldwide,and previous studies have shown that,in addition to affecting the eyes,it also causes abnormalities in the brain.However,it is not yet clear how the primary visual cortex(V1)is altered in glaucoma.This study used DBA/2J mice as a model for spontaneous secondary glaucoma.The aim of the study was to compare the electrophysiological and histomorphological chara cteristics of neurons in the V1between 9-month-old DBA/2J mice and age-matched C57BL/6J mice.We conducted single-unit recordings in the V1 of light-anesthetized mice to measure the visually induced responses,including single-unit spiking and gamma band oscillations.The morphology of layerⅡ/Ⅲneurons was determined by neuronal nuclear antigen staining and Nissl staining of brain tissue sections.Eighty-seven neurons from eight DBA/2J mice and eighty-one neurons from eight C57BL/6J mice were examined.Compared with the C57BL/6J group,V1 neurons in the DBA/2J group exhibited weaker visual tuning and impaired spatial summation.Moreove r,fewer neuro ns were observed in the V1 of DBA/2J mice compared with C57BL/6J mice.These findings suggest that DBA/2J mice have fewer neurons in the VI compared with C57BL/6J mice,and that these neurons have impaired visual tuning.Our findings provide a better understanding of the pathological changes that occur in V1 neuron function and morphology in the DBA/2J mouse model.This study might offer some innovative perspectives regarding the treatment of glaucoma.
基金National Natural Science Foundation of China(Grant Nos.62335006,62022032,62275065,and 61875047)Key Laboratory of Opto-Electronic Information Acquisition and Manipulation(Anhui University),Ministry of Education(Grant No.OEIAM202202)Fundamental Research Funds for the Central Universities(Grant No.HIT.OCEF.2023011).
文摘A highly sensitive light-induced thermoelectric spectroscopy(LITES)sensor based on a multi-pass cell(MPC)with dense spot pattern and a novel quartz tuning fork(QTF)with low resonance frequency is reported in this manuscript.An erbi-um-doped fiber amplifier(EDFA)was employed to amplify the output optical power so that the signal level was further enhanced.The optical path length(OPL)and the ratio of optical path length to volume(RLV)of the MPC is 37.7 m and 13.8 cm^(-2),respectively.A commercial QTF and a self-designed trapezoidal-tip QTF with low frequency of 9461.83 Hz were used as the detectors of the sensor,respectively.The target gas selected to test the performance of the system was acetylene(C2H2).When the optical power was constant at 1000 mW,the minimum detection limit(MDL)of the C2H2-LITES sensor can be achieved 48.3 ppb when using the commercial QTF and 24.6 ppb when using the trapezoid-al-tip QTF.An improvement of the detection performance by a factor of 1.96 was achieved after replacing the commer-cial QTF with the trapezoidal-tip QTF.
文摘Fraud of credit cards is a major issue for financial organizations and individuals.As fraudulent actions become more complex,a demand for better fraud detection systems is rising.Deep learning approaches have shown promise in several fields,including detecting credit card fraud.However,the efficacy of these models is heavily dependent on the careful selection of appropriate hyperparameters.This paper introduces models that integrate deep learning models with hyperparameter tuning techniques to learn the patterns and relationships within credit card transaction data,thereby improving fraud detection.Three deep learning models:AutoEncoder(AE),Convolution Neural Network(CNN),and Long Short-Term Memory(LSTM)are proposed to investigate how hyperparameter adjustment impacts the efficacy of deep learning models used to identify credit card fraud.The experiments conducted on a European credit card fraud dataset using different hyperparameters and three deep learning models demonstrate that the proposed models achieve a tradeoff between detection rate and precision,leading these models to be effective in accurately predicting credit card fraud.The results demonstrate that LSTM significantly outperformed AE and CNN in terms of accuracy(99.2%),detection rate(93.3%),and area under the curve(96.3%).These proposed models have surpassed those of existing studies and are expected to make a significant contribution to the field of credit card fraud detection.
基金This research was funded by the Natural Science Research Project of Higher Education Institutions in Anhui Province(Grant No.2022AH040045)the Anhui Provincial Natural Science Foundation(Grant No.2008085QE245)the Project of Science and Technology Plan of Department of Housing and Urban-Rural Development of Anhui Province(Grant No.2021-YF22).
文摘In order to improve the seismic performance of adjacent buildings,two types of tuned inerter damper(TID)damping systems for adjacent buildings are proposed,which are composed of springs,inerter devices and dampers in serial or in parallel.The dynamic equations of TID adjacent building damping systems were derived,and the H2 norm criterion was used to optimize and adjust them,so that the system had the optimum damping performance under white noise random excitation.Taking TID frequency ratio and damping ratio as optimization parameters,the optimum analytical solutions of the displacement frequency response of the undamped structure under white noise excitation were obtained.The results showed that compared with the classic TMD,TID could obtain a better damping effect in the adjacent buildings.Comparing the TIDs composed of serial or parallel,it was found that the parallel TIDs had more significant advantages in controlling the peak displacement frequency response,while the H2 norm of the displacement frequency response of the damping system under the coupling of serial TID was smaller.Taking the adjacent building composed of two ten-story frame structures as an example,the displacement and energy collection time history analysis of the adjacent building coupled with the optimum design parameter TIDs were carried out.It was found that TID had a better damping effect in the full-time range compared with the classic TMD.This paper also studied the potential power of TID in adjacent buildings,which can be converted into available power resources during earthquakes.
文摘Sentence classification is the process of categorizing a sentence based on the context of the sentence.Sentence categorization requires more semantic highlights than other tasks,such as dependence parsing,which requires more syntactic elements.Most existing strategies focus on the general semantics of a conversation without involving the context of the sentence,recognizing the progress and comparing impacts.An ensemble pre-trained language model was taken up here to classify the conversation sentences from the conversation corpus.The conversational sentences are classified into four categories:information,question,directive,and commission.These classification label sequences are for analyzing the conversation progress and predicting the pecking order of the conversation.Ensemble of Bidirectional Encoder for Representation of Transformer(BERT),Robustly Optimized BERT pretraining Approach(RoBERTa),Generative Pre-Trained Transformer(GPT),DistilBERT and Generalized Autoregressive Pretraining for Language Understanding(XLNet)models are trained on conversation corpus with hyperparameters.Hyperparameter tuning approach is carried out for better performance on sentence classification.This Ensemble of Pre-trained Language Models with a Hyperparameter Tuning(EPLM-HT)system is trained on an annotated conversation dataset.The proposed approach outperformed compared to the base BERT,GPT,DistilBERT and XLNet transformer models.The proposed ensemble model with the fine-tuned parameters achieved an F1_score of 0.88.
文摘Hyperparameter tuning is a key step in developing high-performing machine learning models, but searching large hyperparameter spaces requires extensive computation using standard sequential methods. This work analyzes the performance gains from parallel versus sequential hyperparameter optimization. Using scikit-learn’s Randomized SearchCV, this project tuned a Random Forest classifier for fake news detection via randomized grid search. Setting n_jobs to -1 enabled full parallelization across CPU cores. Results show the parallel implementation achieved over 5× faster CPU times and 3× faster total run times compared to sequential tuning. However, test accuracy slightly dropped from 99.26% sequentially to 99.15% with parallelism, indicating a trade-off between evaluation efficiency and model performance. Still, the significant computational gains allow more extensive hyperparameter exploration within reasonable timeframes, outweighing the small accuracy decrease. Further analysis could better quantify this trade-off across different models, tuning techniques, tasks, and hardware.
文摘超高层建筑是现代城市建设的重要标志之一,其高度已经超过了传统建筑的极限。然而,随着建筑高度不断增加,地震的破坏力也越来越强,超高层建筑面临着更加严峻的安全挑战。因此,研究超高层建筑防震支撑系统技术非常重要,Tuned Mass Damper(TMD)是一种被广泛研究和应用的超高层建筑防震支撑系统技术,TMD最初是在20世纪60年代提出的,最早应用于桥梁上,后来,TMD被引入到建筑领域,并得到广泛的应用。通过精确调节质量、阻尼和弹性等参数来削弱地震引起的建筑物减震效应,从而减少了建筑物因地震造成的损害和崩塌的风险.
基金supported by Korea Institute for Advancement of Technology(KIAT)grant fundedthe Korea Government(MOTIE)(P0012724,The Competency Development Program for Industry Specialist)the Soonchunhyang University Research Fund.
文摘Human Activity Recognition(HAR)has been made simple in recent years,thanks to recent advancements made in Artificial Intelligence(AI)techni-ques.These techniques are applied in several areas like security,surveillance,healthcare,human-robot interaction,and entertainment.Since wearable sensor-based HAR system includes in-built sensors,human activities can be categorized based on sensor values.Further,it can also be employed in other applications such as gait diagnosis,observation of children/adult’s cognitive nature,stroke-patient hospital direction,Epilepsy and Parkinson’s disease examination,etc.Recently-developed Artificial Intelligence(AI)techniques,especially Deep Learning(DL)models can be deployed to accomplish effective outcomes on HAR process.With this motivation,the current research paper focuses on designing Intelligent Hyperparameter Tuned Deep Learning-based HAR(IHPTDL-HAR)technique in healthcare environment.The proposed IHPTDL-HAR technique aims at recogniz-ing the human actions in healthcare environment and helps the patients in mana-ging their healthcare service.In addition,the presented model makes use of Hierarchical Clustering(HC)-based outlier detection technique to remove the out-liers.IHPTDL-HAR technique incorporates DL-based Deep Belief Network(DBN)model to recognize the activities of users.Moreover,Harris Hawks Opti-mization(HHO)algorithm is used for hyperparameter tuning of DBN model.Finally,a comprehensive experimental analysis was conducted upon benchmark dataset and the results were examined under different aspects.The experimental results demonstrate that the proposed IHPTDL-HAR technique is a superior per-former compared to other recent techniques under different measures.
文摘Human pose estimation(HPE)is a procedure for determining the structure of the body pose and it is considered a challenging issue in the computer vision(CV)communities.HPE finds its applications in several fields namely activity recognition and human-computer interface.Despite the benefits of HPE,it is still a challenging process due to the variations in visual appearances,lighting,occlusions,dimensionality,etc.To resolve these issues,this paper presents a squirrel search optimization with a deep convolutional neural network for HPE(SSDCNN-HPE)technique.The major intention of the SSDCNN-HPE technique is to identify the human pose accurately and efficiently.Primarily,the video frame conversion process is performed and pre-processing takes place via bilateral filtering-based noise removal process.Then,the EfficientNet model is applied to identify the body points of a person with no problem constraints.Besides,the hyperparameter tuning of the EfficientNet model takes place by the use of the squirrel search algorithm(SSA).In the final stage,the multiclass support vector machine(M-SVM)technique was utilized for the identification and classification of human poses.The design of bilateral filtering followed by SSA based EfficientNetmodel for HPE depicts the novelty of the work.To demonstrate the enhanced outcomes of the SSDCNN-HPE approach,a series of simulations are executed.The experimental results reported the betterment of the SSDCNN-HPE system over the recent existing techniques in terms of different measures.
基金supported by the Technology Development Program of MSS(No.S3033853).
文摘Recently,computation offloading has become an effective method for overcoming the constraint of a mobile device(MD)using computationintensivemobile and offloading delay-sensitive application tasks to the remote cloud-based data center.Smart city benefitted from offloading to edge point.Consider a mobile edge computing(MEC)network in multiple regions.They comprise N MDs and many access points,in which everyMDhasM independent real-time tasks.This study designs a new Task Offloading and Resource Allocation in IoT-based MEC using Deep Learning with Seagull Optimization(TORA-DLSGO)algorithm.The proposed TORA-DLSGO technique addresses the resource management issue in the MEC server,which enables an optimum offloading decision to minimize the system cost.In addition,an objective function is derived based on minimizing energy consumption subject to the latency requirements and restricted resources.The TORA-DLSGO technique uses the deep belief network(DBN)model for optimum offloading decision-making.Finally,the SGO algorithm is used for the parameter tuning of the DBN model.The simulation results exemplify that the TORA-DLSGO technique outperformed the existing model in reducing client overhead in the MEC systems with a maximum reward of 0.8967.
文摘Remote sensing image(RSI)classifier roles a vital play in earth observation technology utilizing Remote sensing(RS)data are extremely exploited from both military and civil fields.More recently,as novel DL approaches develop,techniques for RSI classifiers with DL have attained important breakthroughs,providing a new opportunity for the research and development of RSI classifiers.This study introduces an Improved Slime Mould Optimization with a graph convolutional network for the hyperspectral remote sensing image classification(ISMOGCN-HRSC)model.The ISMOGCN-HRSC model majorly concentrates on identifying and classifying distinct kinds of RSIs.In the presented ISMOGCN-HRSC model,the synergic deep learning(SDL)model is exploited to produce feature vectors.The GCN model is utilized for image classification purposes to identify the proper class labels of the RSIs.The ISMO algorithm is used to enhance the classification efficiency of the GCN method,which is derived by integrating chaotic concepts into the SMO algorithm.The experimental assessment of the ISMOGCN-HRSC method is tested using a benchmark dataset.