The H∞ output feedback control problem for uncertain discrete-time switched systems is reasearclled. A new characterization of stability and H∞ performance for the switched system under arbitrary switching is obtain...The H∞ output feedback control problem for uncertain discrete-time switched systems is reasearclled. A new characterization of stability and H∞ performance for the switched system under arbitrary switching is obtained by using switched Lyapunov function. Then, based on the characterization, a linear matrix inequality （LMI） approach is developed to design a switched output feedback controller which guarantees the stability and H∞ performance of the closed-loop system. A numerical example is presented to demonstrate the application of the proposed method.展开更多
In this paper,the robust stability issue of switched uncertain multidelay systems resulting from actuator failures is considered.Based on the average dwell time approach,a set of suitable switching signals is designed...In this paper,the robust stability issue of switched uncertain multidelay systems resulting from actuator failures is considered.Based on the average dwell time approach,a set of suitable switching signals is designed by using the total activation time ratio between the stable subsystem and the unstable one.It is first proven that the resulting closed-loop system is robustly exponentially stable for some allowable upper bound of delays if the nominal system with zero delay is exponentially stable under these switching laws.Particularly,the maximal upper bound of delays can be obtained from the linear matrix inequalities.At last,the effectiveness of the proposed method is demonstrated by a simulation example.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
A robust dissipative control problem for a class of It-type stochastic systems is discussed with Markovian jumping parameters and time-varying delay. A memoryless state feedback dissipative controller is developed bas...A robust dissipative control problem for a class of It-type stochastic systems is discussed with Markovian jumping parameters and time-varying delay. A memoryless state feedback dissipative controller is developed based on Lyapunov-Krasovskii functional approach such that the closed-loop system is robustly stochastically stable and weakly delay-dependent （RSSWDD） and strictly （Q, S, R）-dissipative. The sufficient condition on the existence of state feedback dissipative controller is presented by linear matrix inequality （LMI）. And the desired controller can be concluded as solving a set of LMI. Finally, a numerical example is provided to demonstrate the effectiveness of the proposed approach.展开更多
This paper deals with the robust stabilization problem for a class of nonlinear systems with structural uncertainty. Based on robust control Lyapunov function, a sufficient and necessary condition for a function to be...This paper deals with the robust stabilization problem for a class of nonlinear systems with structural uncertainty. Based on robust control Lyapunov function, a sufficient and necessary condition for a function to be a robust control Lyapunov function is given. From this condition, simply sufficient condition for the robust stabilization （robust practical stabilization） is deduced. Moreover, if the equilibrium of the closed-loop system is unique, the existence of such a robust control Lyapunnv function will also imply robustly globally asymptotical stabilization. Then a continuous state feedback law can be constructed explicitly. The simulation shows the effectiveness of the method.展开更多
This note deals with stabilization of uncertain linear neutral delay systems. A new stabilization scheme is presented. Using new Lyapunov-Krasovskii functionals, less conservative stabilization conditions are derived ...This note deals with stabilization of uncertain linear neutral delay systems. A new stabilization scheme is presented. Using new Lyapunov-Krasovskii functionals, less conservative stabilization conditions are derived for such systems based on linear matrix inequalities （LMI）. The results are illustrated using a numerical example.展开更多
针对低资源语言缺少标签数据,而无法使用现有成熟的深度学习方法进行命名实体识别(NER)的问题,提出基于句级别对抗生成网络(GAN)的跨语言NER模型——SLGAN-XLM-R(Sentence Level GAN Based on XLM-R)。首先,使用源语言的标签数据在预训...针对低资源语言缺少标签数据,而无法使用现有成熟的深度学习方法进行命名实体识别(NER)的问题,提出基于句级别对抗生成网络(GAN)的跨语言NER模型——SLGAN-XLM-R(Sentence Level GAN Based on XLM-R)。首先,使用源语言的标签数据在预训练模型XLM-R (XLM-Robustly optimized BERT pretraining approach)的基础上训练NER模型;同时,结合目标语言的无标签数据对XLM-R模型的嵌入层进行语言对抗训练;然后,使用NER模型来预测目标语言无标签数据的软标签;最后,混合源语言与目标语言的标签数据,以对模型进行二次微调来得到最终的NER模型。在CoNLL2002和CoNLL2003两个数据集的英语、德语、西班牙语、荷兰语四种语言上的实验结果表明,以英语作为源语言时,SLGAN-XLM-R模型在德语、西班牙语、荷兰语测试集上的F1值分别为72.70%、79.42%、80.03%,相较于直接在XLM-R模型上进行微调分别提升了5.38、5.38、3.05个百分点。展开更多
基金the National Natural Science Foundation of China (60574083)the Scientific Research Foundation for the Returned Overseas Chinese Scholars (SRF for ROCS),State Education Ministry of China.
文摘The H∞ output feedback control problem for uncertain discrete-time switched systems is reasearclled. A new characterization of stability and H∞ performance for the switched system under arbitrary switching is obtained by using switched Lyapunov function. Then, based on the characterization, a linear matrix inequality （LMI） approach is developed to design a switched output feedback controller which guarantees the stability and H∞ performance of the closed-loop system. A numerical example is presented to demonstrate the application of the proposed method.
基金supported by the National Basic Research Program of China (No. 2007CB714006)the National Natural Science Foundation(No. 61074020)
文摘In this paper,the robust stability issue of switched uncertain multidelay systems resulting from actuator failures is considered.Based on the average dwell time approach,a set of suitable switching signals is designed by using the total activation time ratio between the stable subsystem and the unstable one.It is first proven that the resulting closed-loop system is robustly exponentially stable for some allowable upper bound of delays if the nominal system with zero delay is exponentially stable under these switching laws.Particularly,the maximal upper bound of delays can be obtained from the linear matrix inequalities.At last,the effectiveness of the proposed method is demonstrated by a simulation example.
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.
基金supported in part by the National Natural Science Foundation of China (60874045 60904030)+1 种基金the Foundation of the Education Bureau of Jiangsu Province (09KJB510019)the Natural Science Foundation of Jiangsu Province (BK2009184)
文摘A robust dissipative control problem for a class of It-type stochastic systems is discussed with Markovian jumping parameters and time-varying delay. A memoryless state feedback dissipative controller is developed based on Lyapunov-Krasovskii functional approach such that the closed-loop system is robustly stochastically stable and weakly delay-dependent （RSSWDD） and strictly （Q, S, R）-dissipative. The sufficient condition on the existence of state feedback dissipative controller is presented by linear matrix inequality （LMI）. And the desired controller can be concluded as solving a set of LMI. Finally, a numerical example is provided to demonstrate the effectiveness of the proposed approach.
基金Sponsored by the Natural Science Foundation of Zhejiang Province in China(Grant No. Y105141).
文摘This paper deals with the robust stabilization problem for a class of nonlinear systems with structural uncertainty. Based on robust control Lyapunov function, a sufficient and necessary condition for a function to be a robust control Lyapunov function is given. From this condition, simply sufficient condition for the robust stabilization （robust practical stabilization） is deduced. Moreover, if the equilibrium of the closed-loop system is unique, the existence of such a robust control Lyapunnv function will also imply robustly globally asymptotical stabilization. Then a continuous state feedback law can be constructed explicitly. The simulation shows the effectiveness of the method.
文摘This note deals with stabilization of uncertain linear neutral delay systems. A new stabilization scheme is presented. Using new Lyapunov-Krasovskii functionals, less conservative stabilization conditions are derived for such systems based on linear matrix inequalities （LMI）. The results are illustrated using a numerical example.
文摘针对低资源语言缺少标签数据,而无法使用现有成熟的深度学习方法进行命名实体识别(NER)的问题,提出基于句级别对抗生成网络(GAN)的跨语言NER模型——SLGAN-XLM-R(Sentence Level GAN Based on XLM-R)。首先,使用源语言的标签数据在预训练模型XLM-R (XLM-Robustly optimized BERT pretraining approach)的基础上训练NER模型;同时,结合目标语言的无标签数据对XLM-R模型的嵌入层进行语言对抗训练;然后,使用NER模型来预测目标语言无标签数据的软标签;最后,混合源语言与目标语言的标签数据,以对模型进行二次微调来得到最终的NER模型。在CoNLL2002和CoNLL2003两个数据集的英语、德语、西班牙语、荷兰语四种语言上的实验结果表明,以英语作为源语言时,SLGAN-XLM-R模型在德语、西班牙语、荷兰语测试集上的F1值分别为72.70%、79.42%、80.03%,相较于直接在XLM-R模型上进行微调分别提升了5.38、5.38、3.05个百分点。
基金National Natural Science Foundation of P. R. China (60374037)the Program for New Century Excellent Talents in Universities of P.R.China (NCET)the Specialized Research Fund for the Doctoral Program of Higher Education of China (20050055013)
