Sleep disturbances are among the most prevalent neuropsychiatric symptoms in individuals who have recovered from severe acute respiratory syndrome coronavirus 2 infections.Previous studies have demonstrated abnormal b...Sleep disturbances are among the most prevalent neuropsychiatric symptoms in individuals who have recovered from severe acute respiratory syndrome coronavirus 2 infections.Previous studies have demonstrated abnormal brain structures in patients with sleep disturbances who have recovered from coronavirus disease 2019(COVID-19).However,neuroimaging studies on sleep disturbances caused by COVID-19 are scarce,and existing studies have primarily focused on the long-term effects of the virus,with minimal acute phase data.As a result,little is known about the pathophysiology of sleep disturbances in the acute phase of COVID-19.To address this issue,we designed a longitudinal study to investigate whether alterations in brain structure occur during the acute phase of infection,and verified the results using 3-month follow-up data.A total of 26 COVID-19 patients with sleep disturbances(aged 51.5±13.57 years,8 women and 18 men),27 COVID-19 patients without sleep disturbances(aged 47.33±15.98 years,9 women and 18 men),and 31 age-and gender-matched healthy controls(aged 49.19±17.51 years,9 women and 22 men)were included in this study.Eleven COVID-19 patients with sleep disturbances were included in a longitudinal analysis.We found that COVID-19 patients with sleep disturbances exhibited brain structural changes in almost all brain lobes.The cortical thicknesses of the left pars opercularis and left precuneus were significantly negatively correlated with Pittsburgh Sleep Quality Index scores.Additionally,we observed changes in the volume of the hippocampus and its subfield regions in COVID-19 patients compared with the healthy controls.The 3-month follow-up data revealed indices of altered cerebral structure(cortical thickness,cortical grey matter volume,and cortical surface area)in the frontal-parietal cortex compared with the baseline in COVID-19 patients with sleep disturbances.Our findings indicate that the sleep disturbances patients had altered morphology in the cortical and hippocampal structures during the acute phase of infection and persistent changes in cortical regions at 3 months post-infection.These data improve our understanding of the pathophysiology of sleep disturbances caused by COVID-19.展开更多
Neuromorphic computing extends beyond sequential processing modalities and outperforms traditional von Neumann architectures in implementing more complicated tasks,e.g.,pattern processing,image recognition,and decisio...Neuromorphic computing extends beyond sequential processing modalities and outperforms traditional von Neumann architectures in implementing more complicated tasks,e.g.,pattern processing,image recognition,and decision making.It features parallel interconnected neural networks,high fault tolerance,robustness,autonomous learning capability,and ultralow energy dissipation.The algorithms of artificial neural network(ANN)have also been widely used because of their facile self-organization and self-learning capabilities,which mimic those of the human brain.To some extent,ANN reflects several basic functions of the human brain and can be efficiently integrated into neuromorphic devices to perform neuromorphic computations.This review highlights recent advances in neuromorphic devices assisted by machine learning algorithms.First,the basic structure of simple neuron models inspired by biological neurons and the information processing in simple neural networks are particularly discussed.Second,the fabrication and research progress of neuromorphic devices are presented regarding to materials and structures.Furthermore,the fabrication of neuromorphic devices,including stand-alone neuromorphic devices,neuromorphic device arrays,and integrated neuromorphic systems,is discussed and demonstrated with reference to some respective studies.The applications of neuromorphic devices assisted by machine learning algorithms in different fields are categorized and investigated.Finally,perspectives,suggestions,and potential solutions to the current challenges of neuromorphic devices are provided.展开更多
In the big data era,the surge in network traffic volume poses challenges for network management and cybersecurity.Network Traffic Classification(NTC)employs deep learning to categorize traffic data,aiding security and...In the big data era,the surge in network traffic volume poses challenges for network management and cybersecurity.Network Traffic Classification(NTC)employs deep learning to categorize traffic data,aiding security and analysis systems as Intrusion Detection Systems(IDS)and Intrusion Prevention Systems(IPS).However,current NTC methods,based on isolated network simulations,usually fail to adapt to new protocols and applications and ignore the effects of network conditions and user behavior on traffic patterns.To improve network traffic management insights,federated learning frameworks have been proposed to aggregate diverse traffic data for collaborative model training.This approach faces challenges like data integrity,label noise,packet loss,and skewed data distributions.While label noise can be mitigated through the use of sophisticated traffic labeling tools,other issues such as packet loss and skewed data distributions encountered in Network Packet Brokers(NPB)can severely impede the efficacy of federated learning algorithms.In this paper,we introduced the Robust Traffic Classifier with Federated Contrastive Learning(FC-RTC),combining federated and contrastive learning methods.Using the Supcon-Loss function from contrastive learning,FC-RTC distinguishes between similar and dissimilar samples.Training by sample pairs,FC-RTC effectively updates when receiving corrupted traffic data with packet loss or disorder.In cases of sample imbalance,contrastive loss functions for similar samples reduce model bias towards higher proportion data.By addressing uneven data distribution and packet loss,our system enhances its capability to adapt and perform accurately in real-world network traffic analysis,meeting the specific demands of this complex field.展开更多
With the arrival of the era of artificial intelligence(AI)and big data,the explosive growth of data has raised higher demands on computer hardware and systems.Neuromorphic techniques inspired by biological nervous sys...With the arrival of the era of artificial intelligence(AI)and big data,the explosive growth of data has raised higher demands on computer hardware and systems.Neuromorphic techniques inspired by biological nervous systems are expected to be one of the approaches to breaking the von Neumann bottleneck.Piezotronic neuromorphic devices modulate electrical transport characteristics by piezopotential and directly associate external mechanical motion with electrical output signals in an active manner,with the capability to sense/store/process information of external stimuli.In this review,we have presented the piezotronic neuromorphic devices(which are classified into strain-gated piezotronic transistors and piezoelectric nanogenerator-gated field effect transistors based on device structure)and discussed their operating mechanisms and related manufacture techniques.Secondly,we summarized the research progress of piezotronic neuromorphic devices in recent years and provided a detailed discussion on multifunctional applications,including bionic sensing,information storage,logic computing,and electrical/optical artificial synapses.Finally,in the context of future development,challenges,and perspectives,we have discussed how to modulate novel neuromorphic devices with piezotronic effects more effectively.It is believed that the piezotronic neuromorphic devices have great potential for the next generation of interactive sensation/memory/computation to facilitate the development of the Internet of Things,AI,biomedical engineering,etc.展开更多
Contact electrification-activated triboelectric potential offers an efficient route to tuning the transport properties in semiconductor devices through electrolyte dielectrics,i.e.,triboiontronics.Organic electrochemi...Contact electrification-activated triboelectric potential offers an efficient route to tuning the transport properties in semiconductor devices through electrolyte dielectrics,i.e.,triboiontronics.Organic electrochemical transistors(OECTs)make more effective use of ion injection in the electrolyte dielectrics by changing the doping state of the semiconductor channel.However,the mainstream flexible/wearable electronics and OECT-based devices are usually modulated by electrical signals and constructed in conventional geometry,which lack direct and efficient interaction between the external environment and functional electronic devices.Here,we demonstrate a fiber-shaped triboiontronic electrochemical transistor with good electrical performances,including a current on/off ratio as high as≈1286 with off-current at~nA level,the average threshold displacements(D_(th))of 0.3 mm,the subthreshold swing corresponding to displacement(SS_(D))at 1.6 mm/dec,and excellent flexibility and durability.The proposed triboiontronic electrochemical transistor has great potential to be used in flexible,functional,and smart self-powered electronic textile.展开更多
Two-dimensional(2D)tribotronic devices have been successfully involved in electromechanical modulation for channel conductance and applied in intelligent sensing system,touch screen,and logic gates.Ambipolar transisto...Two-dimensional(2D)tribotronic devices have been successfully involved in electromechanical modulation for channel conductance and applied in intelligent sensing system,touch screen,and logic gates.Ambipolar transistors and corresponding complementary inverters based on one type of semiconductors are highly promising due to the facile fabrication process and readily tunable polarity.Here,we demonstrate an ambipolar tribotronic transistor of molybdenum ditelluride(MoTe_(2)),which shows typical ambipolar transport properties modulated by triboelectric potential.It is comprised of a MoTe_(2)transistor and a lateral sliding triboelectric nanogenerator(TENG).The induced triboelectric potential by Maxwell’s displacement current(a driving force for TENG)can readily modulate the transport properties of both electrons and holes in MoTe_(2)channel and effectively drive the transistor.High performance tribotronic properties have been achieved,including low cutoff current below 1 pA·μm^(−1)and high current on/off ratio of~103 for holes and electrons dominated transports.The working mechanism on how to achieve tribotronic ambipolarity is discussed in detail.A complementary tribotronic inverter based on single flake of MoTe_(2)is also demonstrated with low power consumption and high stability.This work presents an active approach to efficiently modulate semiconductor devices and logic circuits based on 2D materials through external mechanical signal,which has great potential in human–machine interaction,intelligent sensor,and other wearable devices.展开更多
The appearance of covid-19 has ravaged the global and triggered a economic recession.This essay aims to predict the macroeconomy after the pendamic shock.We start with an analysis of the correlation between covid and ...The appearance of covid-19 has ravaged the global and triggered a economic recession.This essay aims to predict the macroeconomy after the pendamic shock.We start with an analysis of the correlation between covid and economic mobility,and then try to make predictions about GDP,mainly using some machine learning models.Several machine learning models are built to forecast and then we estimate their performances.In detail,first,we will try to predict US GDP using all models.After estimating their results,we are able to choose the best model among all.Then we use this model to forecast Italy’s GDP in order to make sure its ability at a larger scales.To make comparison,we also use traditional VAR model to predict and get its performance.The conclusion of this paper shows that the LSTM model performs the best among all the machine learning models.However,compared with the traditional VAR autoregressive model,there was still a gap of 2.6 times.展开更多
基金supported by grants from Major Project of Science and Technology of Guangxi Zhuang Autonomous Region,No.Guike-AA22096018(to JY)Guangxi Key Research and Development Program,No.AB22080053(to DD)+6 种基金Major Project of Science and Technology of Guangxi Zhuang Autonomous Region,No.Guike-AA23023004(to MZ)the National Natural Science Foundation of China,Nos.82260021(to MZ),82060315(to DD)the Natural Science Foundation of Guangxi Zhuang Autonomous Region,No.2021GXNSFBA220007(to GD)Clinical Research Center For Medical Imaging in Hunan Province,No.2020SK4001(to JL)Key Emergency Project of Pneumonia Epidemic of Novel Coronavirus Infection in Hunan Province,No.2020SK3006(to JL)Science and Technology Innovation Program of Hunan Province,No.2021RC4016(to JL)Key Project of the Natural Science Foundation of Hunan Province,No.2024JJ3041(to JL).
文摘Sleep disturbances are among the most prevalent neuropsychiatric symptoms in individuals who have recovered from severe acute respiratory syndrome coronavirus 2 infections.Previous studies have demonstrated abnormal brain structures in patients with sleep disturbances who have recovered from coronavirus disease 2019(COVID-19).However,neuroimaging studies on sleep disturbances caused by COVID-19 are scarce,and existing studies have primarily focused on the long-term effects of the virus,with minimal acute phase data.As a result,little is known about the pathophysiology of sleep disturbances in the acute phase of COVID-19.To address this issue,we designed a longitudinal study to investigate whether alterations in brain structure occur during the acute phase of infection,and verified the results using 3-month follow-up data.A total of 26 COVID-19 patients with sleep disturbances(aged 51.5±13.57 years,8 women and 18 men),27 COVID-19 patients without sleep disturbances(aged 47.33±15.98 years,9 women and 18 men),and 31 age-and gender-matched healthy controls(aged 49.19±17.51 years,9 women and 22 men)were included in this study.Eleven COVID-19 patients with sleep disturbances were included in a longitudinal analysis.We found that COVID-19 patients with sleep disturbances exhibited brain structural changes in almost all brain lobes.The cortical thicknesses of the left pars opercularis and left precuneus were significantly negatively correlated with Pittsburgh Sleep Quality Index scores.Additionally,we observed changes in the volume of the hippocampus and its subfield regions in COVID-19 patients compared with the healthy controls.The 3-month follow-up data revealed indices of altered cerebral structure(cortical thickness,cortical grey matter volume,and cortical surface area)in the frontal-parietal cortex compared with the baseline in COVID-19 patients with sleep disturbances.Our findings indicate that the sleep disturbances patients had altered morphology in the cortical and hippocampal structures during the acute phase of infection and persistent changes in cortical regions at 3 months post-infection.These data improve our understanding of the pathophysiology of sleep disturbances caused by COVID-19.
基金financially supported by the National Natural Science Foundation of China(No.52073031)the National Key Research and Development Program of China(Nos.2023YFB3208102,2021YFB3200304)+4 种基金the China National Postdoctoral Program for Innovative Talents(No.BX2021302)the Beijing Nova Program(Nos.Z191100001119047,Z211100002121148)the Fundamental Research Funds for the Central Universities(No.E0EG6801X2)the‘Hundred Talents Program’of the Chinese Academy of Sciencesthe BrainLink program funded by the MSIT through the NRF of Korea(No.RS-2023-00237308).
文摘Neuromorphic computing extends beyond sequential processing modalities and outperforms traditional von Neumann architectures in implementing more complicated tasks,e.g.,pattern processing,image recognition,and decision making.It features parallel interconnected neural networks,high fault tolerance,robustness,autonomous learning capability,and ultralow energy dissipation.The algorithms of artificial neural network(ANN)have also been widely used because of their facile self-organization and self-learning capabilities,which mimic those of the human brain.To some extent,ANN reflects several basic functions of the human brain and can be efficiently integrated into neuromorphic devices to perform neuromorphic computations.This review highlights recent advances in neuromorphic devices assisted by machine learning algorithms.First,the basic structure of simple neuron models inspired by biological neurons and the information processing in simple neural networks are particularly discussed.Second,the fabrication and research progress of neuromorphic devices are presented regarding to materials and structures.Furthermore,the fabrication of neuromorphic devices,including stand-alone neuromorphic devices,neuromorphic device arrays,and integrated neuromorphic systems,is discussed and demonstrated with reference to some respective studies.The applications of neuromorphic devices assisted by machine learning algorithms in different fields are categorized and investigated.Finally,perspectives,suggestions,and potential solutions to the current challenges of neuromorphic devices are provided.
基金supported by the Joint Funds of the National Natural Science Foundation of China under grant No.U22B2025.
文摘In the big data era,the surge in network traffic volume poses challenges for network management and cybersecurity.Network Traffic Classification(NTC)employs deep learning to categorize traffic data,aiding security and analysis systems as Intrusion Detection Systems(IDS)and Intrusion Prevention Systems(IPS).However,current NTC methods,based on isolated network simulations,usually fail to adapt to new protocols and applications and ignore the effects of network conditions and user behavior on traffic patterns.To improve network traffic management insights,federated learning frameworks have been proposed to aggregate diverse traffic data for collaborative model training.This approach faces challenges like data integrity,label noise,packet loss,and skewed data distributions.While label noise can be mitigated through the use of sophisticated traffic labeling tools,other issues such as packet loss and skewed data distributions encountered in Network Packet Brokers(NPB)can severely impede the efficacy of federated learning algorithms.In this paper,we introduced the Robust Traffic Classifier with Federated Contrastive Learning(FC-RTC),combining federated and contrastive learning methods.Using the Supcon-Loss function from contrastive learning,FC-RTC distinguishes between similar and dissimilar samples.Training by sample pairs,FC-RTC effectively updates when receiving corrupted traffic data with packet loss or disorder.In cases of sample imbalance,contrastive loss functions for similar samples reduce model bias towards higher proportion data.By addressing uneven data distribution and packet loss,our system enhances its capability to adapt and perform accurately in real-world network traffic analysis,meeting the specific demands of this complex field.
基金financially supported by the National Natural Science Foundation of China(52073031,22008151)the National Key Research and Development Program of China(2021YFB3200304)+2 种基金Beijing Nova Program(Z211100002121148)Fundamental Research Funds for the Central Universities(E0EG6801X2)the‘Hundred Talents Program’of the Chinese Academy of Sciences。
文摘With the arrival of the era of artificial intelligence(AI)and big data,the explosive growth of data has raised higher demands on computer hardware and systems.Neuromorphic techniques inspired by biological nervous systems are expected to be one of the approaches to breaking the von Neumann bottleneck.Piezotronic neuromorphic devices modulate electrical transport characteristics by piezopotential and directly associate external mechanical motion with electrical output signals in an active manner,with the capability to sense/store/process information of external stimuli.In this review,we have presented the piezotronic neuromorphic devices(which are classified into strain-gated piezotronic transistors and piezoelectric nanogenerator-gated field effect transistors based on device structure)and discussed their operating mechanisms and related manufacture techniques.Secondly,we summarized the research progress of piezotronic neuromorphic devices in recent years and provided a detailed discussion on multifunctional applications,including bionic sensing,information storage,logic computing,and electrical/optical artificial synapses.Finally,in the context of future development,challenges,and perspectives,we have discussed how to modulate novel neuromorphic devices with piezotronic effects more effectively.It is believed that the piezotronic neuromorphic devices have great potential for the next generation of interactive sensation/memory/computation to facilitate the development of the Internet of Things,AI,biomedical engineering,etc.
基金supported by the National Key Research and Development Program of China(2016YFA0202703,2016YFA0202701)the Fundamental Research Funds for the Central Universities(E0EG6801X2)+2 种基金the National Natural Science Foundation of China(52073031,51605034,and 51711540300)the Beijing Nova Program(Z191100001119047)the“Hundred Talents Program”of the Chinese Academy of Science.
文摘Contact electrification-activated triboelectric potential offers an efficient route to tuning the transport properties in semiconductor devices through electrolyte dielectrics,i.e.,triboiontronics.Organic electrochemical transistors(OECTs)make more effective use of ion injection in the electrolyte dielectrics by changing the doping state of the semiconductor channel.However,the mainstream flexible/wearable electronics and OECT-based devices are usually modulated by electrical signals and constructed in conventional geometry,which lack direct and efficient interaction between the external environment and functional electronic devices.Here,we demonstrate a fiber-shaped triboiontronic electrochemical transistor with good electrical performances,including a current on/off ratio as high as≈1286 with off-current at~nA level,the average threshold displacements(D_(th))of 0.3 mm,the subthreshold swing corresponding to displacement(SS_(D))at 1.6 mm/dec,and excellent flexibility and durability.The proposed triboiontronic electrochemical transistor has great potential to be used in flexible,functional,and smart self-powered electronic textile.
基金financially supported by the National Key Research and Development Program of China(No.2021YFB3200304)the National Natural Science Foundation of China(No.52073031)+2 种基金the Beijing Nova Program(Nos.Z191100001119047 and Z211100002121148)the Fundamental Research Funds for the Central Universities(No.E0EG6801X2)the“Hundred Talents Program”of the Chinese Academy of Sciences.
文摘Two-dimensional(2D)tribotronic devices have been successfully involved in electromechanical modulation for channel conductance and applied in intelligent sensing system,touch screen,and logic gates.Ambipolar transistors and corresponding complementary inverters based on one type of semiconductors are highly promising due to the facile fabrication process and readily tunable polarity.Here,we demonstrate an ambipolar tribotronic transistor of molybdenum ditelluride(MoTe_(2)),which shows typical ambipolar transport properties modulated by triboelectric potential.It is comprised of a MoTe_(2)transistor and a lateral sliding triboelectric nanogenerator(TENG).The induced triboelectric potential by Maxwell’s displacement current(a driving force for TENG)can readily modulate the transport properties of both electrons and holes in MoTe_(2)channel and effectively drive the transistor.High performance tribotronic properties have been achieved,including low cutoff current below 1 pA·μm^(−1)and high current on/off ratio of~103 for holes and electrons dominated transports.The working mechanism on how to achieve tribotronic ambipolarity is discussed in detail.A complementary tribotronic inverter based on single flake of MoTe_(2)is also demonstrated with low power consumption and high stability.This work presents an active approach to efficiently modulate semiconductor devices and logic circuits based on 2D materials through external mechanical signal,which has great potential in human–machine interaction,intelligent sensor,and other wearable devices.
文摘The appearance of covid-19 has ravaged the global and triggered a economic recession.This essay aims to predict the macroeconomy after the pendamic shock.We start with an analysis of the correlation between covid and economic mobility,and then try to make predictions about GDP,mainly using some machine learning models.Several machine learning models are built to forecast and then we estimate their performances.In detail,first,we will try to predict US GDP using all models.After estimating their results,we are able to choose the best model among all.Then we use this model to forecast Italy’s GDP in order to make sure its ability at a larger scales.To make comparison,we also use traditional VAR model to predict and get its performance.The conclusion of this paper shows that the LSTM model performs the best among all the machine learning models.However,compared with the traditional VAR autoregressive model,there was still a gap of 2.6 times.
