Low pressure chemical vapor deposition(LPCVD) is one of the most important processes during semiconductor manufacturing.However,the spatial distribution of internal temperature and extremely few samples makes it hard ...Low pressure chemical vapor deposition(LPCVD) is one of the most important processes during semiconductor manufacturing.However,the spatial distribution of internal temperature and extremely few samples makes it hard to build a good-quality model of this batch process.Besides,due to the properties of this process,the reliability of the model must be taken into consideration when optimizing the MVs.In this work,an optimal design strategy based on the self-learning Gaussian process model(GPM) is proposed to control this kind of spatial batch process.The GPM is utilized as the internal model to predict the thicknesses of thin films on all spatial-distributed wafers using the limited data.Unlike the conventional model based design,the uncertainties of predictions provided by GPM are taken into consideration to guide the optimal design of manipulated variables so that the designing can be more prudent Besides,the GPM is also actively enhanced using as little data as possible based on the predictive uncertainties.The effectiveness of the proposed strategy is successfully demonstrated in an LPCVD process.展开更多
The contribution of this work is twofold： （1） a multimodality prediction method of chaotic time series with the Gaussian process mixture （GPM） model is proposed, which employs a divide and conquer strategy. It au...The contribution of this work is twofold： （1） a multimodality prediction method of chaotic time series with the Gaussian process mixture （GPM） model is proposed, which employs a divide and conquer strategy. It automatically divides the chaotic time series into multiple modalities with different extrinsic patterns and intrinsic characteristics, and thus can more precisely fit the chaotic time series. （2） An effective sparse hard-cut expec- tation maximization （SHC-EM） learning algorithm for the GPM model is proposed to improve the prediction performance. SHO-EM replaces a large learning sample set with fewer pseudo inputs, accelerating model learning based on these pseudo inputs. Experiments on Lorenz and Chua time series demonstrate that the proposed method yields not only accurate multimodality prediction, but also the prediction confidence interval SHC-EM outperforms the traditional variational 1earning in terms of both prediction accuracy and speed. In addition, SHC-EM is more robust and insusceptible to noise than variational learning.展开更多
Design and operation optimization of centrifugal compressor are always based on an accurate prediction model, however, due to the short time operation and lack of data information, it is difficult to get an accurate p...Design and operation optimization of centrifugal compressor are always based on an accurate prediction model, however, due to the short time operation and lack of data information, it is difficult to get an accurate prediction model of a new centrifugal compressor in time. This paper applies an improved fast model migration method(FMM method) to develop the model of the new centrifugal compressor. The method adapts a Gaussian Process(GP) model from an old centrifugal compressor to fit a new and similar centrifugal compressor, and the adaptation is conducted by a scale-bias adjustment migration technology. In order to obtain the better estimated parameters of migration, Bayesian method, which takes the prior knowledge into consideration, is used in the sequential experiment. The approach is validated by a specific simulation bench. The results indicate that the applied approach can achieve a better prediction precision with fewer data of the new centrifugal compressor compared to pure GP method, and can model the new centrifugal compressor rapidly.展开更多
A single-channel electroencephalography(EEG)device,despite being widely accepted due to convenience,ease of deployment and suitability for use in complex environments,typically poses a great challenge for reactive bra...A single-channel electroencephalography(EEG)device,despite being widely accepted due to convenience,ease of deployment and suitability for use in complex environments,typically poses a great challenge for reactive brain-computer interface(BCI)applications particularly when a continuous command from users is desired to run a motorized actuator with different speed profiles.In this study,a combination of an inconspicuous visual stimulus and voluntary eyeblinks along with a machine learning-based decoder is considered as a new reactive BCI paradigm to increase the degree of freedom and minimize mismatches between the intended dynamic command and transmitted control signal.The proposed decoder is constructed based on Gaussian Process model(GPM)which is a nonparametric Bayesian approach that has the advantages of being able to operate on small datasets and providing measurements of uncertainty on predictions.To evaluate the effectiveness of the proposed method,the GPM is compared against other competitive techniques which include k-Nearest Neighbors,linear discriminant analysis,support vector machine,ensemble learning and neural network.Results demonstrate that a significant improvement can be achieved via the GPM approach with average accuracy reaching over 96%and mean absolute error of no greater than 0.8 cm/s.In addition,the analysis reveals that while the performances of other existing methods deteriorate with a certain type of stimulus due to signal drifts resulting from the voluntary eyeblinks,the proposed GPM exhibits consistent performance across all stimuli considered,thereby manifesting its generalization capability and making it a more suitable option for dynamic commands with a single-channel EEG-controlled actuator.展开更多
The dynamic soft sensor based on a single Gaussian process regression(GPR) model has been developed in fermentation processes.However,limitations of single regression models,for multiphase/multimode fermentation proce...The dynamic soft sensor based on a single Gaussian process regression(GPR) model has been developed in fermentation processes.However,limitations of single regression models,for multiphase/multimode fermentation processes,may result in large prediction errors and complexity of the soft sensor.Therefore,a dynamic soft sensor based on Gaussian mixture regression(GMR) was proposed to overcome the problems.Two structure parameters,the number of Gaussian components and the order of the model,are crucial to the soft sensor model.To achieve a simple and effective soft sensor,an iterative strategy was proposed to optimize the two structure parameters synchronously.For the aim of comparisons,the proposed dynamic GMR soft sensor and the existing dynamic GPR soft sensor were both investigated to estimate biomass concentration in a Penicillin simulation process and an industrial Erythromycin fermentation process.Results show that the proposed dynamic GMR soft sensor has higher prediction accuracy and is more suitable for dynamic multiphase/multimode fermentation processes.展开更多
The metastable retained austenite(RA)plays a significant role in the excellent mechanical performance of quenching and partitioning(Q&P)steels,while the volume fraction of RA(V_(RA))is challengeable to directly pr...The metastable retained austenite(RA)plays a significant role in the excellent mechanical performance of quenching and partitioning(Q&P)steels,while the volume fraction of RA(V_(RA))is challengeable to directly predict due to the complicated relationships between the chemical composition and process(like quenching temperature(Qr)).A Gaussian process regression model in machine learning was developed to predict V_(RA),and the model accuracy was further improved by introducing a metallurgical parameter of martensite fraction(fo)to accurately predict V_(RA) in Q&P steels.The developed machine learning model combined with Bayesian global optimization can serve as another selection strategy for the quenching temperature,and this strategy is very effcient as it found the"optimum"Qr with the maximum V_(RA) using only seven consecutive iterations.The benchmark experiment also reveals that the developed machine learning model predicts V_(RA) more accurately than the popular constrained carbon equilibrium thermodynamic model,even better than a thermo-kinetic quenching-partitioning-tempering-local equilibrium model.展开更多
Voice conversion algorithm aims to provide high level of similarity to the target voice with an acceptable level of quality.The main object of this paper was to build a nonlinear relationship between the parameters fo...Voice conversion algorithm aims to provide high level of similarity to the target voice with an acceptable level of quality.The main object of this paper was to build a nonlinear relationship between the parameters for the acoustical features of source and target speaker using Non-Linear Canonical Correlation Analysis(NLCCA) based on jointed Gaussian mixture model.Speaker indi-viduality transformation was achieved mainly by altering vocal tract characteristics represented by Line Spectral Frequencies(LSF).To obtain the transformed speech which sounded more like the target voices,prosody modification is involved through residual prediction.Both objective and subjective evaluations were conducted.The experimental results demonstrated that our proposed algorithm was effective and outperformed the conventional conversion method utilized by the Minimum Mean Square Error(MMSE) estimation.展开更多
轨迹跟踪是无人驾驶控制系统中至关重要的功能之一。车辆动力学模型对轨迹跟踪性能有显著影响,但是存在模型复杂度和求解效率之间的矛盾,在非线性工况下无法满足轨迹跟踪精度要求,为此提出基于高斯过程回归(Gaussian Process Regression...轨迹跟踪是无人驾驶控制系统中至关重要的功能之一。车辆动力学模型对轨迹跟踪性能有显著影响,但是存在模型复杂度和求解效率之间的矛盾,在非线性工况下无法满足轨迹跟踪精度要求,为此提出基于高斯过程回归(Gaussian Process Regression,GPR)的模型预测控制(Model Predictive Control,MPC)方法。使用简单模型从而确保求解效率,通过GPR对车辆模型补偿从而提高轨迹跟踪性能。提出基于单轨动力学模型的车辆状态融合估计方法,获得GPR误差补偿模型;构建轨迹跟踪问题模型,推导GPR误差补偿模型在预测时域的迭代方程,对预测时域内的车辆状态进行动态补偿,实现轨迹跟踪控制;通过搭建实车验证平台开展典型工况试验验证,与无补偿MPC方法进行对比。研究结果表明,新方法轨迹跟踪精度得到明显提升,轨迹跟踪横向误差和航向误差分别降低了33.3%和27.9%,同时还兼顾了车辆舒适性的提升,侧向加速度和横摆角速度均值分别下降了17.1%和21.7%。展开更多
Efficient experiment design is of great significance for the validation of simulation model with high nonlinearity and large input space.Excessive validation experiment raises the cost while insufficient test increase...Efficient experiment design is of great significance for the validation of simulation model with high nonlinearity and large input space.Excessive validation experiment raises the cost while insufficient test increases the risks of accepting an invalid model.In this paper,an adaptive sequential experiment design method combining global exploration criterion and local exploitation criterion is proposed.The exploration criterion utilizes discrepancy metric to improve the space-filling property of the design points while the exploitation criterion employs the leave one out error to discover informative points.To avoid the clustering of samples in the local region,an adaptive weight updating approach is provided to maintain the balance between exploration and exploitation.Besides,the credibility distribution function characterizing the relationship between the input and result credibility is introduced to support the model validation experiment design.Finally,six benchmark problems and an engineering case are applied to examine the performance of the proposed method.The experiments indicate that the proposed method achieves satisfactory performance for function approximation in accuracy and convergence.展开更多
目的训练多种机器学习模型用于听性脑干反应(auditory brainstem response,ABR)波形的自动识别,并确定准确率最高的模型,使ABR自动识别技术更好地应用于临床实践。方法选取2021年6月至2022年6月北京清华长庚医院收治的100例听力正常和...目的训练多种机器学习模型用于听性脑干反应(auditory brainstem response,ABR)波形的自动识别,并确定准确率最高的模型,使ABR自动识别技术更好地应用于临床实践。方法选取2021年6月至2022年6月北京清华长庚医院收治的100例听力正常和伴有听力损伤人群的受试者(200耳)为研究对象,根据年龄和听力水平将受试者分为组1(年龄18~59岁,500、1000、2000、4000 Hz频率平均听阈≤25 dB HL)、组2(年龄≥60岁,500、1000、2000、4000 Hz频率平均听阈≤25 dB HL)、组3(年龄18~59岁,500、1000、2000、4000 Hz频率平均听阈>25 dB HL)、组4(年龄≥60岁,500、1000、2000、4000 Hz频率平均听阈>25 dB HL),每组25例。收集受试者纯音测听和ABR数据,提取ABR信号时域和频域特征,与受试者年龄、性别、纯音听阈,刺激声强度以及原始信号序列拼接得到特征向量。分别使用逻辑回归、支持向量机分类、伯努利朴素贝叶斯分类、高斯朴素贝叶斯分类、高斯过程分类、决策树、随机森林、表格网络、轻量化梯度提升框架、极致梯度提升框架和局部级联集成。等机器学习模型对ABR波形进行识别训练,并对整体数据和分组数据分别计算不同模型下波形识别的准确率。结果高斯过程分类模型的整体准确率达到了94.89%,超过了其他机器学习模型。其中95.62%为<60岁听力正常受试者、92.19%为≥60岁听力正常受试者、92.92%为<60岁伴有听力损失受试者、92.50%为≥60岁且伴有听力损失受试者。结论机器学习技术在ABR波形的自动识别方面具有良好的应用前景,高斯过程分类模型优于其他机器学习模型。展开更多
基金Supported by the National High Technology Research and Development Program of China(2014AA041803)the National Natural Science Foundation of China(61320106009)
文摘Low pressure chemical vapor deposition(LPCVD) is one of the most important processes during semiconductor manufacturing.However,the spatial distribution of internal temperature and extremely few samples makes it hard to build a good-quality model of this batch process.Besides,due to the properties of this process,the reliability of the model must be taken into consideration when optimizing the MVs.In this work,an optimal design strategy based on the self-learning Gaussian process model(GPM) is proposed to control this kind of spatial batch process.The GPM is utilized as the internal model to predict the thicknesses of thin films on all spatial-distributed wafers using the limited data.Unlike the conventional model based design,the uncertainties of predictions provided by GPM are taken into consideration to guide the optimal design of manipulated variables so that the designing can be more prudent Besides,the GPM is also actively enhanced using as little data as possible based on the predictive uncertainties.The effectiveness of the proposed strategy is successfully demonstrated in an LPCVD process.
基金Supported by the National Natural Science Foundation of China under Grant No 60972106the China Postdoctoral Science Foundation under Grant No 2014M561053+1 种基金the Humanity and Social Science Foundation of Ministry of Education of China under Grant No 15YJA630108the Hebei Province Natural Science Foundation under Grant No E2016202341
文摘The contribution of this work is twofold： （1） a multimodality prediction method of chaotic time series with the Gaussian process mixture （GPM） model is proposed, which employs a divide and conquer strategy. It automatically divides the chaotic time series into multiple modalities with different extrinsic patterns and intrinsic characteristics, and thus can more precisely fit the chaotic time series. （2） An effective sparse hard-cut expec- tation maximization （SHC-EM） learning algorithm for the GPM model is proposed to improve the prediction performance. SHO-EM replaces a large learning sample set with fewer pseudo inputs, accelerating model learning based on these pseudo inputs. Experiments on Lorenz and Chua time series demonstrate that the proposed method yields not only accurate multimodality prediction, but also the prediction confidence interval SHC-EM outperforms the traditional variational 1earning in terms of both prediction accuracy and speed. In addition, SHC-EM is more robust and insusceptible to noise than variational learning.
基金supported by the Fundamental Research Funds for the Central Universities(Grant No.2015QNA65)
文摘Design and operation optimization of centrifugal compressor are always based on an accurate prediction model, however, due to the short time operation and lack of data information, it is difficult to get an accurate prediction model of a new centrifugal compressor in time. This paper applies an improved fast model migration method(FMM method) to develop the model of the new centrifugal compressor. The method adapts a Gaussian Process(GP) model from an old centrifugal compressor to fit a new and similar centrifugal compressor, and the adaptation is conducted by a scale-bias adjustment migration technology. In order to obtain the better estimated parameters of migration, Bayesian method, which takes the prior knowledge into consideration, is used in the sequential experiment. The approach is validated by a specific simulation bench. The results indicate that the applied approach can achieve a better prediction precision with fewer data of the new centrifugal compressor compared to pure GP method, and can model the new centrifugal compressor rapidly.
基金This work was supported by the Ministry of Higher Education Malaysia for Fundamental Research Grant Scheme with Project Code:FRGS/1/2021/TK0/USM/02/18.
文摘A single-channel electroencephalography(EEG)device,despite being widely accepted due to convenience,ease of deployment and suitability for use in complex environments,typically poses a great challenge for reactive brain-computer interface(BCI)applications particularly when a continuous command from users is desired to run a motorized actuator with different speed profiles.In this study,a combination of an inconspicuous visual stimulus and voluntary eyeblinks along with a machine learning-based decoder is considered as a new reactive BCI paradigm to increase the degree of freedom and minimize mismatches between the intended dynamic command and transmitted control signal.The proposed decoder is constructed based on Gaussian Process model(GPM)which is a nonparametric Bayesian approach that has the advantages of being able to operate on small datasets and providing measurements of uncertainty on predictions.To evaluate the effectiveness of the proposed method,the GPM is compared against other competitive techniques which include k-Nearest Neighbors,linear discriminant analysis,support vector machine,ensemble learning and neural network.Results demonstrate that a significant improvement can be achieved via the GPM approach with average accuracy reaching over 96%and mean absolute error of no greater than 0.8 cm/s.In addition,the analysis reveals that while the performances of other existing methods deteriorate with a certain type of stimulus due to signal drifts resulting from the voluntary eyeblinks,the proposed GPM exhibits consistent performance across all stimuli considered,thereby manifesting its generalization capability and making it a more suitable option for dynamic commands with a single-channel EEG-controlled actuator.
基金Supported by the Natural Science Foundation of Jiangsu Province of China(BK20130531)the Priority Academic Program Development of Jiangsu Higher Education Institutions(PAPD[2011]6)Jiangsu Government Scholarship
文摘The dynamic soft sensor based on a single Gaussian process regression(GPR) model has been developed in fermentation processes.However,limitations of single regression models,for multiphase/multimode fermentation processes,may result in large prediction errors and complexity of the soft sensor.Therefore,a dynamic soft sensor based on Gaussian mixture regression(GMR) was proposed to overcome the problems.Two structure parameters,the number of Gaussian components and the order of the model,are crucial to the soft sensor model.To achieve a simple and effective soft sensor,an iterative strategy was proposed to optimize the two structure parameters synchronously.For the aim of comparisons,the proposed dynamic GMR soft sensor and the existing dynamic GPR soft sensor were both investigated to estimate biomass concentration in a Penicillin simulation process and an industrial Erythromycin fermentation process.Results show that the proposed dynamic GMR soft sensor has higher prediction accuracy and is more suitable for dynamic multiphase/multimode fermentation processes.
基金The authors acknowledge financial support from the National Natural Science Foundation of China(Grant Nos.51771114 and 51371117).
文摘The metastable retained austenite(RA)plays a significant role in the excellent mechanical performance of quenching and partitioning(Q&P)steels,while the volume fraction of RA(V_(RA))is challengeable to directly predict due to the complicated relationships between the chemical composition and process(like quenching temperature(Qr)).A Gaussian process regression model in machine learning was developed to predict V_(RA),and the model accuracy was further improved by introducing a metallurgical parameter of martensite fraction(fo)to accurately predict V_(RA) in Q&P steels.The developed machine learning model combined with Bayesian global optimization can serve as another selection strategy for the quenching temperature,and this strategy is very effcient as it found the"optimum"Qr with the maximum V_(RA) using only seven consecutive iterations.The benchmark experiment also reveals that the developed machine learning model predicts V_(RA) more accurately than the popular constrained carbon equilibrium thermodynamic model,even better than a thermo-kinetic quenching-partitioning-tempering-local equilibrium model.
基金Supported by the National High Technology Research and Development Program of China (863 Program,No.2006AA010102)
文摘Voice conversion algorithm aims to provide high level of similarity to the target voice with an acceptable level of quality.The main object of this paper was to build a nonlinear relationship between the parameters for the acoustical features of source and target speaker using Non-Linear Canonical Correlation Analysis(NLCCA) based on jointed Gaussian mixture model.Speaker indi-viduality transformation was achieved mainly by altering vocal tract characteristics represented by Line Spectral Frequencies(LSF).To obtain the transformed speech which sounded more like the target voices,prosody modification is involved through residual prediction.Both objective and subjective evaluations were conducted.The experimental results demonstrated that our proposed algorithm was effective and outperformed the conventional conversion method utilized by the Minimum Mean Square Error(MMSE) estimation.
文摘轨迹跟踪是无人驾驶控制系统中至关重要的功能之一。车辆动力学模型对轨迹跟踪性能有显著影响,但是存在模型复杂度和求解效率之间的矛盾,在非线性工况下无法满足轨迹跟踪精度要求,为此提出基于高斯过程回归(Gaussian Process Regression,GPR)的模型预测控制(Model Predictive Control,MPC)方法。使用简单模型从而确保求解效率,通过GPR对车辆模型补偿从而提高轨迹跟踪性能。提出基于单轨动力学模型的车辆状态融合估计方法,获得GPR误差补偿模型;构建轨迹跟踪问题模型,推导GPR误差补偿模型在预测时域的迭代方程,对预测时域内的车辆状态进行动态补偿,实现轨迹跟踪控制;通过搭建实车验证平台开展典型工况试验验证,与无补偿MPC方法进行对比。研究结果表明,新方法轨迹跟踪精度得到明显提升,轨迹跟踪横向误差和航向误差分别降低了33.3%和27.9%,同时还兼顾了车辆舒适性的提升,侧向加速度和横摆角速度均值分别下降了17.1%和21.7%。
基金supported by the National Natural Science Foundation of China(No.61627810)。
文摘Efficient experiment design is of great significance for the validation of simulation model with high nonlinearity and large input space.Excessive validation experiment raises the cost while insufficient test increases the risks of accepting an invalid model.In this paper,an adaptive sequential experiment design method combining global exploration criterion and local exploitation criterion is proposed.The exploration criterion utilizes discrepancy metric to improve the space-filling property of the design points while the exploitation criterion employs the leave one out error to discover informative points.To avoid the clustering of samples in the local region,an adaptive weight updating approach is provided to maintain the balance between exploration and exploitation.Besides,the credibility distribution function characterizing the relationship between the input and result credibility is introduced to support the model validation experiment design.Finally,six benchmark problems and an engineering case are applied to examine the performance of the proposed method.The experiments indicate that the proposed method achieves satisfactory performance for function approximation in accuracy and convergence.
文摘目的训练多种机器学习模型用于听性脑干反应(auditory brainstem response,ABR)波形的自动识别,并确定准确率最高的模型,使ABR自动识别技术更好地应用于临床实践。方法选取2021年6月至2022年6月北京清华长庚医院收治的100例听力正常和伴有听力损伤人群的受试者(200耳)为研究对象,根据年龄和听力水平将受试者分为组1(年龄18~59岁,500、1000、2000、4000 Hz频率平均听阈≤25 dB HL)、组2(年龄≥60岁,500、1000、2000、4000 Hz频率平均听阈≤25 dB HL)、组3(年龄18~59岁,500、1000、2000、4000 Hz频率平均听阈>25 dB HL)、组4(年龄≥60岁,500、1000、2000、4000 Hz频率平均听阈>25 dB HL),每组25例。收集受试者纯音测听和ABR数据,提取ABR信号时域和频域特征,与受试者年龄、性别、纯音听阈,刺激声强度以及原始信号序列拼接得到特征向量。分别使用逻辑回归、支持向量机分类、伯努利朴素贝叶斯分类、高斯朴素贝叶斯分类、高斯过程分类、决策树、随机森林、表格网络、轻量化梯度提升框架、极致梯度提升框架和局部级联集成。等机器学习模型对ABR波形进行识别训练,并对整体数据和分组数据分别计算不同模型下波形识别的准确率。结果高斯过程分类模型的整体准确率达到了94.89%,超过了其他机器学习模型。其中95.62%为<60岁听力正常受试者、92.19%为≥60岁听力正常受试者、92.92%为<60岁伴有听力损失受试者、92.50%为≥60岁且伴有听力损失受试者。结论机器学习技术在ABR波形的自动识别方面具有良好的应用前景,高斯过程分类模型优于其他机器学习模型。
