As a category of recurrent neural networks,echo state networks(ESNs)have been the topic of in-depth investigations and extensive applications in a diverse array of fields,with spectacular triumphs achieved.Nevertheles...As a category of recurrent neural networks,echo state networks(ESNs)have been the topic of in-depth investigations and extensive applications in a diverse array of fields,with spectacular triumphs achieved.Nevertheless,the traditional ESN and the majority of its variants are devised in the light of the second-order statistical information of data(e.g.,variance and covariance),while more information is neglected.In the context of information theoretic learning,correntropy demonstrates the capacity to grab more information from data.Therefore,under the guidelines of the maximum correntropy criterion,this paper proposes a correntropy-based echo state network(CESN)in which the first-order and higher-order information of data is captured,promoting robustness to noise.Furthermore,an incremental learning algorithm for the CESN is presented,which has the expertise to update the CESN when new data arrives,eliminating the need to retrain the network from scratch.Finally,experiments on benchmark problems and comparisons with existing works are provided to verify the effectiveness and superiority of the proposed CESN.展开更多
Recent decades have witnessed a trend that the echo state network(ESN)is widely utilized in field of time series prediction due to its powerful computational abilities.However,most of the existing research on ESN is c...Recent decades have witnessed a trend that the echo state network(ESN)is widely utilized in field of time series prediction due to its powerful computational abilities.However,most of the existing research on ESN is conducted under the assumption that data is free of noise or polluted by the Gaussian noise,which lacks robustness or even fails to solve real-world tasks.This work handles this issue by proposing a probabilistic regularized ESN(PRESN)with robustness guaranteed.Specifically,we design a novel objective function for minimizing both the mean and variance of modeling error,and then a scheme is derived for getting output weights of the PRESN.Furthermore,generalization performance,robustness,and unbiased estimation abilities of the PRESN are revealed by theoretical analyses.Finally,experiments on a benchmark dataset and two real-world datasets are conducted to verify the performance of the proposed PRESN.The source code is publicly available at https://github.com/LongJinlab/probabilistic-regularized-echo-state-network.展开更多
The Chaotic Baseband Wireless Communication System(CBWCS)is expected to eliminate the Inter-Symbol Interference(ISI)caused by multipath propagation by using the optimal decoding threshold that is the sum of the ISI ca...The Chaotic Baseband Wireless Communication System(CBWCS)is expected to eliminate the Inter-Symbol Interference(ISI)caused by multipath propagation by using the optimal decoding threshold that is the sum of the ISI caused by past decoded bits and the ISI caused by future transmitting bits.However,the current technique is only capable of removing partial effects of the ISI,because only past decoded bits are available for the suboptimal decoding threshold calculation.The unavailability of the future information needed for the optimal decoding threshold is an obstacle to further improve the Bit Error Rate(BER)performance.In contrast to the previous method using Echo State Network(ESN)to predict one future bit,the proposed method in this paper predicts the optimal decoding threshold directly using ESN.The proposed ESN-based threshold prediction method simplifies the symbol decoding operation by avoiding the iterative prediction of the output waveform points using ESN and accumulated error caused by the iterative operation.With this approach,the calculation complexity is reduced compared to the previous ESN-based approach.The proposed method achieves better BER performance compared to the previous method.The reason for this superior result is twofold.First,the proposed ESN is capable of using more future symbols information conveyed by the ESN input to obtain more accurate threshold rather than the previous method in which only one future symbol was available.Second,the proposed method here does not need to estimate the channel information using Least Squared(LS)method,which avoids the extra error caused by inaccurate channel information estimation.Simulation results and experiment based on a wireless open-access research platform under a practical wireless channel show the effectiveness and superiority of the proposed method.展开更多
Forecasting stock prices using deep learning models suffers from pro-blems such as low accuracy,slow convergence,and complex network structures.This study developed an echo state network(ESN)model to mitigate such pro...Forecasting stock prices using deep learning models suffers from pro-blems such as low accuracy,slow convergence,and complex network structures.This study developed an echo state network(ESN)model to mitigate such pro-blems.We compared our ESN with a long short-term memory(LSTM)network by forecasting the stock data of Kweichow Moutai,a leading enterprise in China’s liquor industry.By analyzing data for 120,240,and 300 days,we generated fore-cast data for the next 40,80,and 100 days,respectively,using both ESN and LSTM.In terms of accuracy,ESN had the unique advantage of capturing non-linear data.Mean absolute error(MAE)was used to present the accuracy results.The MAEs of the data forecast by ESN were 0.024,0.024,and 0.025,which were,respectively,0.065,0.007,and 0.009 less than those of LSTM.In terms of con-vergence,ESN has a reservoir state-space structure,which makes it perform faster than other models.Root-mean-square error(RMSE)was used to present the con-vergence time.In our experiment,the RMSEs of ESN were 0.22,0.27,and 0.26,which were,respectively,0.08,0.01,and 0.12 less than those of LSTM.In terms of network structure,ESN consists only of input,reservoir,and output spaces,making it a much simpler model than the others.The proposed ESN was found to be an effective model that,compared to others,converges faster,forecasts more accurately,and builds time-series analyses more easily.展开更多
<div style="text-align:justify;"> This paper proposes a prediction method based on improved Echo State Network for COVID-19 nonlinear time series, which improves the Echo State Network from the reservo...<div style="text-align:justify;"> This paper proposes a prediction method based on improved Echo State Network for COVID-19 nonlinear time series, which improves the Echo State Network from the reservoir topology and the output weight matrix, and adopt the ABC (Artificial Bee Colony) algorithm based on crossover and crowding strategy to optimize the parameters. Finally, the proposed method is simulated and the results show that it has stronger prediction ability for COVID-19 nonlinear time series. </div>展开更多
Error accumulation in long-term predictions of chaotic climate systems is caused primarily by the model's high sensitivity to initial conditions and the absence of dynamic adjustment mechanisms,leading to gradual ...Error accumulation in long-term predictions of chaotic climate systems is caused primarily by the model's high sensitivity to initial conditions and the absence of dynamic adjustment mechanisms,leading to gradual forecast divergence.This presents a critical challenge to achieving stable long-term predictions.While current data-driven approaches perform well in short-term forecasting,their accuracy deteriorates significantly over time.To overcome this limitation,we propose an autonomous echo state network with a snow ablation optimizer(AESN-SAO),which significantly improves the adaptability and robustness of data-driven methods under varying initial conditions.This approach not only eliminates the need for manual hyperparameter tuning in traditional AESNs but also effectively mitigates the common issue of initial conditions sensitivity in chaotic climate systems.Furthermore,we introduce a sparse observation insertion mechanism based on the Lyapunov time and valid prediction time(VPT),which enables AESNSAO to correct errors prior to system divergence,effectively extending the prediction horizon.Numerical experiments conducted on the Lorenz-63 and Climate Lorenz-63 systems demonstrate that integrating sparse observations with AESN-SAO approach extends the VPT to approximately 99 Lyapunov times,markedly reducing error accumulation in long-term forecasts.This study provides a reliable and efficient framework for long-term predictions in climate systems with nonlinear and chaotic dynamics,with promising applications in weather forecasting,climate modeling,and disaster risk assessment.展开更多
With the aim to enhance prediction accuracy for nonlinear time series,this paper put forward an improved deep Echo State Network based on reservoir states reconstruction driven by a Self-Normalizing Activation(SNA)fun...With the aim to enhance prediction accuracy for nonlinear time series,this paper put forward an improved deep Echo State Network based on reservoir states reconstruction driven by a Self-Normalizing Activation(SNA)function as the replacement for the traditional Hyperbolic tangent activation function to reduce the model’s sensitivity to hyper-parameters.The Strategy was implemented in a two-state reconstruction process byfirst inputting the time series data to the model separately.Once,the time data passes through the reservoirs and is activated by the SNA activation function,the new state for the reservoirs is created.The state is input to the next layer,and the concatenate states module saves.Pairs of states are selected from the activated multi-layer reservoirs and input into the state reconstruction module.Multiple input states are transformed through the state reconstruction module andfinally saved to the concatenate state module.Two evaluation metrics were used to benchmark against three other ESNs with SNA activation functions to achieve better prediction accuracy.展开更多
Echo state network （ESN） has become one of the most popular recurrent neural networks （RNN） for its good prediction performance of non-linear time series and simple training process. But several problems still pre...Echo state network （ESN） has become one of the most popular recurrent neural networks （RNN） for its good prediction performance of non-linear time series and simple training process. But several problems still prevent ESN from becoming a widely used tool. The most prominent problem is its high complexity with lots of random parameters. Aiming at this problem, a minimum complexity ESN model （MCESN） was proposed. In this paper, we proposed a new wavelet minimum complexity ESN model （WMCESN） to improve the prediction accuracy and increase the practical applicability. Our new model inherits the characters of minimum complexity ESN model using the fixed parameters and simple circle topology. We injected wavelet neurons to replace the original neurons in internal reservoir and designed a wavelet parameter matrix to reduce the computing time. By using different datasets, our new model performed better than the minimum complexity ESN model with normal neurons, but only utilized tiny time cost. We also used our own packets of transmission control protocol （TCP） and user datagram protocol （UDP） dataset to prove that our model can deal with the data packet bit prediction problem well.展开更多
High rigidity twenty-high Sendzimir mills （ZRMs） are widely used for rolling stainless steels, silicon sheets, etc. A ZRM uses a small diameter work roll to produce massive rolling forces. Since a work roll with a s...High rigidity twenty-high Sendzimir mills （ZRMs） are widely used for rolling stainless steels, silicon sheets, etc. A ZRM uses a small diameter work roll to produce massive rolling forces. Since a work roll with a small diameter can be bent easily, strips often have complex shapes with mixed quarter and deep edge waves in the shape of plates. In order to solve this problem, fuzzy neural network controls are generally used for shape： recognition in ZRM control systems. Among various neural network types, the multi-layer perceptron （MLP） is typically used in current ZRMs. However, an MLP causes the loss of a large amount of shape recognition data. To improve the shape recognition per- formance of ZRM control systems, echo state networks （ESNs） are proposed to be used. Through simulation re- sults, it is found that shape recognition performance could be improved using the proposed ESN method.展开更多
With the challenge from services diversity grows greatly,the service-oriented supporting ability is required to current high-speed passive optical network(PON) .Aimed to enhance the quality of service(Qo S) brought by...With the challenge from services diversity grows greatly,the service-oriented supporting ability is required to current high-speed passive optical network(PON) .Aimed to enhance the quality of service(Qo S) brought by diversified-services,this paper proposes an Simplified Echo State Network(SESN) Based Services Awareness scheme in High-Speed PON(Passive Optical Network) .In this proposed scheme,the ring topology is adopted in the reservoir of SESN to reduce the complexity of original Echo State Network,and system dynamics equation is introduced to keep the accuracy of SESN.According to the network architecture of 10G-EPON,a SESN Master is running in the OLT and a number of SESN Agents work in ONUs.The SESN Master plays the main function of service-awareness from the total view of various kinds services in 10G-EPON system,by fully SESN training.Then,the reservoir information of well-trained SESN in OLT will be broadcasted to all ONUs and those SESN Agents working in ONUs are allowed to conducts independent service-awareness function.Thus,resources allocation and transport policy are both determined just only in ONUs.Simulation results show that the proposed mechanism is able to better supporting ability for multiple services.展开更多
Echo state network (ESN), which efficiently models nonlinear dynamic systems, has been proposed as a special form of recurrent neural network. However, most of the proposed ESNs consist of complex reservoir structures...Echo state network (ESN), which efficiently models nonlinear dynamic systems, has been proposed as a special form of recurrent neural network. However, most of the proposed ESNs consist of complex reservoir structures, leading to excessive computational cost. Recently, minimum complexity ESNs were proposed and proved to exhibit high performance and low computational cost. In this paper, we propose a simple deterministic ESN with a loop reservoir, i.e., an ESN with an adjacent-feedback loop reservoir. The novel reservoir is constructed by introducing regular adjacent feedback based on the simplest loop reservoir. Only a single free parameter is tuned, which considerably simplifies the ESN construction. The combination of a simplified reservoir and fewer free parameters provides superior prediction performance. In the benchmark datasets and real-world tasks, our scheme obtains higher prediction accuracy with relatively low complexity, compared to the classic ESN and the minimum complexity ESN. Furthermore, we prove that all the linear ESNs with the simplest loop reservoir possess the same memory capacity, arbitrarily converging to the optimal value.展开更多
With the challenge of great growing of services diversity,service-oriented supporting ability is required by current high-speed passive optical network( PON). Aimed at enhancing the quality of service( Qo S) brought b...With the challenge of great growing of services diversity,service-oriented supporting ability is required by current high-speed passive optical network( PON). Aimed at enhancing the quality of service( Qo S) brought by diversified-services,this study proposes an echo state network( ESN)based multi-service awareness mechanism in 10-Gigabite ethernet passive optical network( 10GEPON). In the proposed approach,distributed architecture is adopted to realize this ESN based multi-service awareness. According to the network architecture of 10G-EPON,where a main ESN is running in OLT and a number of ESN agents works in ONUs. The main-ESN plays the main function of service-awareness from the total view of various kinds of services in 10G-EPON system,by full ESN training. Then,the reservoir information of well-trained ESN in OLT will be broadcasted to all ONUs and those ESN agents working in ONUs are allowed to conduct independent service-awareness function. Thus,resources allocation and transport policy are both determined only in ONUs. Simulation results show that the proposed mechanism is able to better support the ability of multiple services.展开更多
Aimed to enhance the supporting ability for diversified services, this paper proposes a hierarchy echo state network（HESE） based service-awareness（SA）（HESN-SA） mechanism in 10 Gbit/s Ethernet passive optical net...Aimed to enhance the supporting ability for diversified services, this paper proposes a hierarchy echo state network（HESE） based service-awareness（SA）（HESN-SA） mechanism in 10 Gbit/s Ethernet passive optical network（10G-EPON）. In this HESN-SA, hierarchy architecture is adopted to realize echo state network（ESN） classification based SA. According to the network architecture of 10G-EPON, the parent-ESN（p-ESN） module works in the optical line terminal（OLT）, while the sub-ESN（s-ESN） module is embedded in optical network units（ONUs）. Thus, the p-ESN plays the main function of SA with a total view of this system, and s-ESN in each ONU conducts the SA function under the control of p-ESN. Thus, resources allocation and transport policy are both determined by the proposed mechanism through cooperation between OLT and ONUs. Simulation results show that the HESN-SA can improve the supporting ability for multiple services.展开更多
The independent hypothesis between frames in vocal effect(VE) recognition makes it difficult for frame based spectral features to describe the intrinsic temporal correlation and dynamic change information in speech ph...The independent hypothesis between frames in vocal effect(VE) recognition makes it difficult for frame based spectral features to describe the intrinsic temporal correlation and dynamic change information in speech phenomena. A novel VE detection method based on echo state network(ESN) is proposed. The input sequences are mapped into a fixed-dimensionality vector in high dimensional coding space by reservoir of the ESN. Then, radial basis function(RBF) networks are employed to fit the probability density function(pdf) of each VE mode by using the vectors in the high dimensional coding space. Finally, the minimum error rate Bayesian decision is employed to judge the VE mode. The experiments which are conducted on isolated words test set achieve 79.5% average recognition accuracy, and the results show that the proposed method can overcome the defect of the independent hypothesis between frames effectively.展开更多
基金supported in part by the National Natural Science Foundation of China(62176109,62476115)the Fundamental Research Funds for the Central Universities(lzujbky-2023-ey07,lzujbky-2023-it14)+1 种基金the Natural Science Foundation of Gansu Province(24JRRA488)the Supercomputing Center of Lanzhou University
文摘As a category of recurrent neural networks,echo state networks(ESNs)have been the topic of in-depth investigations and extensive applications in a diverse array of fields,with spectacular triumphs achieved.Nevertheless,the traditional ESN and the majority of its variants are devised in the light of the second-order statistical information of data(e.g.,variance and covariance),while more information is neglected.In the context of information theoretic learning,correntropy demonstrates the capacity to grab more information from data.Therefore,under the guidelines of the maximum correntropy criterion,this paper proposes a correntropy-based echo state network(CESN)in which the first-order and higher-order information of data is captured,promoting robustness to noise.Furthermore,an incremental learning algorithm for the CESN is presented,which has the expertise to update the CESN when new data arrives,eliminating the need to retrain the network from scratch.Finally,experiments on benchmark problems and comparisons with existing works are provided to verify the effectiveness and superiority of the proposed CESN.
基金supported in part by the National Natural Science Foundation of China(62176109)the CAAI-Huawei MindSpore Open Fund(CAAIXSJLJJ-2022-020A)+3 种基金the Natural Science Foundation of Gansu Province(21JR7RA531,22JR5RA427,22JR5RA487)the Fundamental Research Funds for the Central Universities(lzujbky-2022-kb12,lzujbky-2022-23)the Science and Technology Project of Chengguan Discrict of Lanzhou(2021-1-2)the Supercomputing Center of Lanzhou University。
文摘Recent decades have witnessed a trend that the echo state network(ESN)is widely utilized in field of time series prediction due to its powerful computational abilities.However,most of the existing research on ESN is conducted under the assumption that data is free of noise or polluted by the Gaussian noise,which lacks robustness or even fails to solve real-world tasks.This work handles this issue by proposing a probabilistic regularized ESN(PRESN)with robustness guaranteed.Specifically,we design a novel objective function for minimizing both the mean and variance of modeling error,and then a scheme is derived for getting output weights of the PRESN.Furthermore,generalization performance,robustness,and unbiased estimation abilities of the PRESN are revealed by theoretical analyses.Finally,experiments on a benchmark dataset and two real-world datasets are conducted to verify the performance of the proposed PRESN.The source code is publicly available at https://github.com/LongJinlab/probabilistic-regularized-echo-state-network.
文摘The Chaotic Baseband Wireless Communication System(CBWCS)is expected to eliminate the Inter-Symbol Interference(ISI)caused by multipath propagation by using the optimal decoding threshold that is the sum of the ISI caused by past decoded bits and the ISI caused by future transmitting bits.However,the current technique is only capable of removing partial effects of the ISI,because only past decoded bits are available for the suboptimal decoding threshold calculation.The unavailability of the future information needed for the optimal decoding threshold is an obstacle to further improve the Bit Error Rate(BER)performance.In contrast to the previous method using Echo State Network(ESN)to predict one future bit,the proposed method in this paper predicts the optimal decoding threshold directly using ESN.The proposed ESN-based threshold prediction method simplifies the symbol decoding operation by avoiding the iterative prediction of the output waveform points using ESN and accumulated error caused by the iterative operation.With this approach,the calculation complexity is reduced compared to the previous ESN-based approach.The proposed method achieves better BER performance compared to the previous method.The reason for this superior result is twofold.First,the proposed ESN is capable of using more future symbols information conveyed by the ESN input to obtain more accurate threshold rather than the previous method in which only one future symbol was available.Second,the proposed method here does not need to estimate the channel information using Least Squared(LS)method,which avoids the extra error caused by inaccurate channel information estimation.Simulation results and experiment based on a wireless open-access research platform under a practical wireless channel show the effectiveness and superiority of the proposed method.
基金supported by the National Natural Science Foundation of China(No.72073041)Open Foundation for the University Innovation Platform in Hunan Province(No.18K103)+2 种基金2011 Collaborative Innovation Center for Development and Utilization of Finance and Economics Big Data Property,Universities of Hunan Province,Open Project(Nos.20181901CRP03,20181901CRP04,20181901CRP05)2020 Hunan Provincial Higher Education Teaching Reform Research Project(Nos.HNJG-2020-1130,HNJG-2020-1124)2020 General Project of Hunan Social Science Fund(No.20B16).
文摘Forecasting stock prices using deep learning models suffers from pro-blems such as low accuracy,slow convergence,and complex network structures.This study developed an echo state network(ESN)model to mitigate such pro-blems.We compared our ESN with a long short-term memory(LSTM)network by forecasting the stock data of Kweichow Moutai,a leading enterprise in China’s liquor industry.By analyzing data for 120,240,and 300 days,we generated fore-cast data for the next 40,80,and 100 days,respectively,using both ESN and LSTM.In terms of accuracy,ESN had the unique advantage of capturing non-linear data.Mean absolute error(MAE)was used to present the accuracy results.The MAEs of the data forecast by ESN were 0.024,0.024,and 0.025,which were,respectively,0.065,0.007,and 0.009 less than those of LSTM.In terms of con-vergence,ESN has a reservoir state-space structure,which makes it perform faster than other models.Root-mean-square error(RMSE)was used to present the con-vergence time.In our experiment,the RMSEs of ESN were 0.22,0.27,and 0.26,which were,respectively,0.08,0.01,and 0.12 less than those of LSTM.In terms of network structure,ESN consists only of input,reservoir,and output spaces,making it a much simpler model than the others.The proposed ESN was found to be an effective model that,compared to others,converges faster,forecasts more accurately,and builds time-series analyses more easily.
文摘<div style="text-align:justify;"> This paper proposes a prediction method based on improved Echo State Network for COVID-19 nonlinear time series, which improves the Echo State Network from the reservoir topology and the output weight matrix, and adopt the ABC (Artificial Bee Colony) algorithm based on crossover and crowding strategy to optimize the parameters. Finally, the proposed method is simulated and the results show that it has stronger prediction ability for COVID-19 nonlinear time series. </div>
基金supported by the National Natural Science Foundation of China(Grant Nos.42430112,42371377)。
文摘Error accumulation in long-term predictions of chaotic climate systems is caused primarily by the model's high sensitivity to initial conditions and the absence of dynamic adjustment mechanisms,leading to gradual forecast divergence.This presents a critical challenge to achieving stable long-term predictions.While current data-driven approaches perform well in short-term forecasting,their accuracy deteriorates significantly over time.To overcome this limitation,we propose an autonomous echo state network with a snow ablation optimizer(AESN-SAO),which significantly improves the adaptability and robustness of data-driven methods under varying initial conditions.This approach not only eliminates the need for manual hyperparameter tuning in traditional AESNs but also effectively mitigates the common issue of initial conditions sensitivity in chaotic climate systems.Furthermore,we introduce a sparse observation insertion mechanism based on the Lyapunov time and valid prediction time(VPT),which enables AESNSAO to correct errors prior to system divergence,effectively extending the prediction horizon.Numerical experiments conducted on the Lorenz-63 and Climate Lorenz-63 systems demonstrate that integrating sparse observations with AESN-SAO approach extends the VPT to approximately 99 Lyapunov times,markedly reducing error accumulation in long-term forecasts.This study provides a reliable and efficient framework for long-term predictions in climate systems with nonlinear and chaotic dynamics,with promising applications in weather forecasting,climate modeling,and disaster risk assessment.
基金supported in part by the National Natural Science Foundation of China under Grant Nos.72171172 and 62088101in part by Shanghai Municipal Science and Technology,China Major Project under Grant No.2021SHZDZX0100+1 种基金in part by Shanghai Research Institute of China Engineering Science and Technology Development Strategy,Strategic Research and Consulting Project,under Grant No.2022-DFZD-33-02in part by Chinese Academy of Engineering,Strategic Research and Consulting Program,under Grant No.2022-XY-100。
文摘With the aim to enhance prediction accuracy for nonlinear time series,this paper put forward an improved deep Echo State Network based on reservoir states reconstruction driven by a Self-Normalizing Activation(SNA)function as the replacement for the traditional Hyperbolic tangent activation function to reduce the model’s sensitivity to hyper-parameters.The Strategy was implemented in a two-state reconstruction process byfirst inputting the time series data to the model separately.Once,the time data passes through the reservoirs and is activated by the SNA activation function,the new state for the reservoirs is created.The state is input to the next layer,and the concatenate states module saves.Pairs of states are selected from the activated multi-layer reservoirs and input into the state reconstruction module.Multiple input states are transformed through the state reconstruction module andfinally saved to the concatenate state module.Two evaluation metrics were used to benchmark against three other ESNs with SNA activation functions to achieve better prediction accuracy.
基金supported by the National Natural Science Foundation of China (61201153)the National Basic Research Program of China (2012CB315805)the National Key Science and Technology Projects (2010ZX03004-002-02)
文摘Echo state network （ESN） has become one of the most popular recurrent neural networks （RNN） for its good prediction performance of non-linear time series and simple training process. But several problems still prevent ESN from becoming a widely used tool. The most prominent problem is its high complexity with lots of random parameters. Aiming at this problem, a minimum complexity ESN model （MCESN） was proposed. In this paper, we proposed a new wavelet minimum complexity ESN model （WMCESN） to improve the prediction accuracy and increase the practical applicability. Our new model inherits the characters of minimum complexity ESN model using the fixed parameters and simple circle topology. We injected wavelet neurons to replace the original neurons in internal reservoir and designed a wavelet parameter matrix to reduce the computing time. By using different datasets, our new model performed better than the minimum complexity ESN model with normal neurons, but only utilized tiny time cost. We also used our own packets of transmission control protocol （TCP） and user datagram protocol （UDP） dataset to prove that our model can deal with the data packet bit prediction problem well.
基金Sponsored by Korea Science and Engineering Foundation(KOSEF)Funded by Korea Government(MEST)(2010-0022521)
文摘High rigidity twenty-high Sendzimir mills （ZRMs） are widely used for rolling stainless steels, silicon sheets, etc. A ZRM uses a small diameter work roll to produce massive rolling forces. Since a work roll with a small diameter can be bent easily, strips often have complex shapes with mixed quarter and deep edge waves in the shape of plates. In order to solve this problem, fuzzy neural network controls are generally used for shape： recognition in ZRM control systems. Among various neural network types, the multi-layer perceptron （MLP） is typically used in current ZRMs. However, an MLP causes the loss of a large amount of shape recognition data. To improve the shape recognition per- formance of ZRM control systems, echo state networks （ESNs） are proposed to be used. Through simulation re- sults, it is found that shape recognition performance could be improved using the proposed ESN method.
基金supported by the Science and Technology Project of State Grid Corporation of China:“Research on the Communication Architecture and Hardware-In-the-Loop Simu-lation of Real-Time Wide-Area Stability Control for Electric Power System”
文摘With the challenge from services diversity grows greatly,the service-oriented supporting ability is required to current high-speed passive optical network(PON) .Aimed to enhance the quality of service(Qo S) brought by diversified-services,this paper proposes an Simplified Echo State Network(SESN) Based Services Awareness scheme in High-Speed PON(Passive Optical Network) .In this proposed scheme,the ring topology is adopted in the reservoir of SESN to reduce the complexity of original Echo State Network,and system dynamics equation is introduced to keep the accuracy of SESN.According to the network architecture of 10G-EPON,a SESN Master is running in the OLT and a number of SESN Agents work in ONUs.The SESN Master plays the main function of service-awareness from the total view of various kinds services in 10G-EPON system,by fully SESN training.Then,the reservoir information of well-trained SESN in OLT will be broadcasted to all ONUs and those SESN Agents working in ONUs are allowed to conducts independent service-awareness function.Thus,resources allocation and transport policy are both determined just only in ONUs.Simulation results show that the proposed mechanism is able to better supporting ability for multiple services.
基金Project supported by the National Basic Research Program (973) of China (No. 2012CB315805)the Fundamental Research Funds for the Central Universities, China (No. 2009RC0124)+1 种基金the National Key Science and Technology Projects, China (No. 2010ZX03004-002-02)the Australian Centre for Broadband Innovation (ACBI)
文摘Echo state network (ESN), which efficiently models nonlinear dynamic systems, has been proposed as a special form of recurrent neural network. However, most of the proposed ESNs consist of complex reservoir structures, leading to excessive computational cost. Recently, minimum complexity ESNs were proposed and proved to exhibit high performance and low computational cost. In this paper, we propose a simple deterministic ESN with a loop reservoir, i.e., an ESN with an adjacent-feedback loop reservoir. The novel reservoir is constructed by introducing regular adjacent feedback based on the simplest loop reservoir. Only a single free parameter is tuned, which considerably simplifies the ESN construction. The combination of a simplified reservoir and fewer free parameters provides superior prediction performance. In the benchmark datasets and real-world tasks, our scheme obtains higher prediction accuracy with relatively low complexity, compared to the classic ESN and the minimum complexity ESN. Furthermore, we prove that all the linear ESNs with the simplest loop reservoir possess the same memory capacity, arbitrarily converging to the optimal value.
基金Supported by the National High Technology Research and Development Programme of China(No.2012AA050804)
文摘With the challenge of great growing of services diversity,service-oriented supporting ability is required by current high-speed passive optical network( PON). Aimed at enhancing the quality of service( Qo S) brought by diversified-services,this study proposes an echo state network( ESN)based multi-service awareness mechanism in 10-Gigabite ethernet passive optical network( 10GEPON). In the proposed approach,distributed architecture is adopted to realize this ESN based multi-service awareness. According to the network architecture of 10G-EPON,where a main ESN is running in OLT and a number of ESN agents works in ONUs. The main-ESN plays the main function of service-awareness from the total view of various kinds of services in 10G-EPON system,by full ESN training. Then,the reservoir information of well-trained ESN in OLT will be broadcasted to all ONUs and those ESN agents working in ONUs are allowed to conduct independent service-awareness function. Thus,resources allocation and transport policy are both determined only in ONUs. Simulation results show that the proposed mechanism is able to better support the ability of multiple services.
基金supported by the Hi-Tech Research and Development Program of China (2012AA050804)
文摘Aimed to enhance the supporting ability for diversified services, this paper proposes a hierarchy echo state network（HESE） based service-awareness（SA）（HESN-SA） mechanism in 10 Gbit/s Ethernet passive optical network（10G-EPON）. In this HESN-SA, hierarchy architecture is adopted to realize echo state network（ESN） classification based SA. According to the network architecture of 10G-EPON, the parent-ESN（p-ESN） module works in the optical line terminal（OLT）, while the sub-ESN（s-ESN） module is embedded in optical network units（ONUs）. Thus, the p-ESN plays the main function of SA with a total view of this system, and s-ESN in each ONU conducts the SA function under the control of p-ESN. Thus, resources allocation and transport policy are both determined by the proposed mechanism through cooperation between OLT and ONUs. Simulation results show that the HESN-SA can improve the supporting ability for multiple services.
基金supported by the National Natural Science Foundation of China (61502150,61300124)the Foundation for University Key Teacher by Henan Province (2015GGJS068)+2 种基金the Fundamental Research Funds for the Universities of Henan Province (NSFRF1616)the Foundation for Scientific and Technological Project of Henan Province (172102210279)the Key Scientific Research Projects of Universities in Henan (19A520004)
文摘The independent hypothesis between frames in vocal effect(VE) recognition makes it difficult for frame based spectral features to describe the intrinsic temporal correlation and dynamic change information in speech phenomena. A novel VE detection method based on echo state network(ESN) is proposed. The input sequences are mapped into a fixed-dimensionality vector in high dimensional coding space by reservoir of the ESN. Then, radial basis function(RBF) networks are employed to fit the probability density function(pdf) of each VE mode by using the vectors in the high dimensional coding space. Finally, the minimum error rate Bayesian decision is employed to judge the VE mode. The experiments which are conducted on isolated words test set achieve 79.5% average recognition accuracy, and the results show that the proposed method can overcome the defect of the independent hypothesis between frames effectively.
