Considering that real communication signals corrupted by noise are generally nonstationary, and timefrequency distributions are especially suitable for the analysis of nonstationary signals, time-frequency distributio...Considering that real communication signals corrupted by noise are generally nonstationary, and timefrequency distributions are especially suitable for the analysis of nonstationary signals, time-frequency distributions are introduced for the modulation classification of communication signals： The extracted time-frequency features have good classification information, and they are insensitive to signal to noise ratio （SNR） variation. According to good classification by the correct rate of a neural network classifier, a multilayer perceptron （MLP） classifier with better generalization, as well as, addition of time-frequency features set for classifying six different modulation types has been proposed. Computer simulations show that the MLP classifier outperforms the decision-theoretic classifier at low SNRs, and the classification experiments for real MPSK signals verify engineering significance of the MLP classifier.展开更多
This paper focuses on some application issues in m.multi-layered perceptrons researches. The following problem areas are discussed: (1) the classification capability of multi-layered perceptrons; (2) theself-configura...This paper focuses on some application issues in m.multi-layered perceptrons researches. The following problem areas are discussed: (1) the classification capability of multi-layered perceptrons; (2) theself-configuration algorithm for facilitating the design of the neural nets' structure;and,finally (3) the application of the fast BP algorithm to speed up the learning procedure. Some experimental results with respect to the application of multi-layered perceptrons as classifier systems in the comprehensive evaluation of Chinese large cities are presented.展开更多
虽然异构计算系统的应用可以加快神经网络参数的处理,但系统功耗也随之剧增。良好的功耗预测方法是异构系统优化功耗和处理多类型工作负载的基础,基于此,通过改进多层感知机-注意力模型,提出一种面向CPU/GPU异构计算系统多类型工作负载...虽然异构计算系统的应用可以加快神经网络参数的处理,但系统功耗也随之剧增。良好的功耗预测方法是异构系统优化功耗和处理多类型工作负载的基础,基于此,通过改进多层感知机-注意力模型,提出一种面向CPU/GPU异构计算系统多类型工作负载的功耗预测算法。首先,考虑服务器功耗与系统特征,建立一种基于特征的工作负载功耗模型;其次,针对现有的功耗预测算法不能解决系统特征与系统功耗之间的长程依赖的问题,提出一种改进的基于多层感知机-注意力模型的功耗预测算法Prophet,该算法改进多层感知机实现各个时刻的系统特征的提取,并使用注意力机制综合这些特征,从而有效解决系统特征与系统功耗之间的长程依赖问题;最后,在实际系统中开展相关实验,将所提算法分别与MLSTM_PM(Power consumption Model based on Multi-layer Long Short-Term Memory)和ENN_PM(Power consumption Model based on Elman Neural Network)等功耗预测算法对比。实验结果表明,Prophet具有较高的预测精准性,与MLSTM_PM算法相比,在工作负载blk、memtest和busspd上将平均相对误差(MRE)分别降低了1.22、1.01和0.93个百分点,并且具有较低的复杂度,表明了所提算法的有效性及可行性。展开更多
文摘Considering that real communication signals corrupted by noise are generally nonstationary, and timefrequency distributions are especially suitable for the analysis of nonstationary signals, time-frequency distributions are introduced for the modulation classification of communication signals： The extracted time-frequency features have good classification information, and they are insensitive to signal to noise ratio （SNR） variation. According to good classification by the correct rate of a neural network classifier, a multilayer perceptron （MLP） classifier with better generalization, as well as, addition of time-frequency features set for classifying six different modulation types has been proposed. Computer simulations show that the MLP classifier outperforms the decision-theoretic classifier at low SNRs, and the classification experiments for real MPSK signals verify engineering significance of the MLP classifier.
文摘This paper focuses on some application issues in m.multi-layered perceptrons researches. The following problem areas are discussed: (1) the classification capability of multi-layered perceptrons; (2) theself-configuration algorithm for facilitating the design of the neural nets' structure;and,finally (3) the application of the fast BP algorithm to speed up the learning procedure. Some experimental results with respect to the application of multi-layered perceptrons as classifier systems in the comprehensive evaluation of Chinese large cities are presented.
文摘虽然异构计算系统的应用可以加快神经网络参数的处理,但系统功耗也随之剧增。良好的功耗预测方法是异构系统优化功耗和处理多类型工作负载的基础,基于此,通过改进多层感知机-注意力模型,提出一种面向CPU/GPU异构计算系统多类型工作负载的功耗预测算法。首先,考虑服务器功耗与系统特征,建立一种基于特征的工作负载功耗模型;其次,针对现有的功耗预测算法不能解决系统特征与系统功耗之间的长程依赖的问题,提出一种改进的基于多层感知机-注意力模型的功耗预测算法Prophet,该算法改进多层感知机实现各个时刻的系统特征的提取,并使用注意力机制综合这些特征,从而有效解决系统特征与系统功耗之间的长程依赖问题;最后,在实际系统中开展相关实验,将所提算法分别与MLSTM_PM(Power consumption Model based on Multi-layer Long Short-Term Memory)和ENN_PM(Power consumption Model based on Elman Neural Network)等功耗预测算法对比。实验结果表明,Prophet具有较高的预测精准性,与MLSTM_PM算法相比,在工作负载blk、memtest和busspd上将平均相对误差(MRE)分别降低了1.22、1.01和0.93个百分点,并且具有较低的复杂度,表明了所提算法的有效性及可行性。
