In this paper, a novel design procedure is proposed for synthesizing high-capacity auto-associative memories based on complex-valued neural networks with real-imaginary-type activation functions and constant delays. S...In this paper, a novel design procedure is proposed for synthesizing high-capacity auto-associative memories based on complex-valued neural networks with real-imaginary-type activation functions and constant delays. Stability criteria dependent on external inputs of neural networks are derived. The designed networks can retrieve the stored patterns by external inputs rather than initial conditions. The derivation can memorize the desired patterns with lower-dimensional neural networks than real-valued neural networks, and eliminate spurious equilibria of complex-valued neural networks. One numerical example is provided to show the effectiveness and superiority of the presented results.展开更多
Memristor is a newly found fourth circuit element for the next generation emerging nonvolatile memory technology. In this paper, design of new type of nonvolatile static random access memory cell is proposed by using ...Memristor is a newly found fourth circuit element for the next generation emerging nonvolatile memory technology. In this paper, design of new type of nonvolatile static random access memory cell is proposed by using a combination of memristor and complemented metal oxide semiconductor. Biolek memristor model and CMOS 180 nm technology are used to form a single cell. By introducing distinct binary logic to avoid safety margin is left for each binary logic output and enables better read/write data integrity. The total power consumption reduces from 0.407 mw (milli-watt) to 0.127 mw which is less than existing memristor based memory cell of the same CMOS technology. Read and write time is also significantly reduced. However, write time is higher than conventional 6T SRAM cell and can be reduced by increasing motion of electron in the memristor. The change of the memristor state is shown by applying piecewise linear input voltage.展开更多
This paper presents the results of a 14-week attention strategy training of 174 college freshmen. It illustrates the promoting function of attention in second language vocabulary acquisition by raising students' expe...This paper presents the results of a 14-week attention strategy training of 174 college freshmen. It illustrates the promoting function of attention in second language vocabulary acquisition by raising students' expectations for new words, by increasing the frequency of exposure to them, by enhancing their perceptual salience, and by increasing the task demand for word study. The results show that enhancing attention in input could promote students' vocabulary acquisition and help them form vocabulary learning strategy suitable for their levels of proficiency.展开更多
由于太阳能和天气变量的随机性和不稳定性,光伏发电具有很高的不确定性,准确的光伏发电功率预测对于光伏电站的短期调度和发电计划的运行至关重要。提出一种基于多模式分解、多通道输入、并联卷积神经网络(parallel convolutional neura...由于太阳能和天气变量的随机性和不稳定性,光伏发电具有很高的不确定性,准确的光伏发电功率预测对于光伏电站的短期调度和发电计划的运行至关重要。提出一种基于多模式分解、多通道输入、并联卷积神经网络(parallel convolutional neural network,PCNN)和双向长短期记忆网络(bi-directional long short term memory,BiLSTM)的组合预测方法,用于不同天气类型的超短期光伏发电功率预测。首先,由相关性分析算法确定辐照度和温度是对光伏发电贡献最大的2个环境变量,并根据环境因素与光伏功率波动特征的关联性将全年数据划分为4类;其次,使用完全集合经验模态分解、奇异谱分解和变分模态分解对辐照度、温度和光伏发电功率进行分解,以降低原始数据的复杂度和非平稳性,实现不同模式模态分量规律互补;最后,建立基于PCNN和BiLSTM的组合预测模型,使用PCNN提取不同的深度特征,并将PCNN输出的特征融合后输入到BiLSTM中,使用BiLSTM建立历史数据之间的时间特征关系,学习历史数据间的正、反向规律,在时空相关性分析的基础上得到最终光伏发电功率预测结果。实验结果表明,提出的组合预测方法在超短期光伏发电功率预测中具有较高的准确性和稳定性,并优于其他深度学习方法。展开更多
基金Project supported by the National Natural Science Foundation of China(Grant Nos.61503338,61573316,61374152,and 11302195)the Natural Science Foundation of Zhejiang Province,China(Grant No.LQ15F030005)
文摘In this paper, a novel design procedure is proposed for synthesizing high-capacity auto-associative memories based on complex-valued neural networks with real-imaginary-type activation functions and constant delays. Stability criteria dependent on external inputs of neural networks are derived. The designed networks can retrieve the stored patterns by external inputs rather than initial conditions. The derivation can memorize the desired patterns with lower-dimensional neural networks than real-valued neural networks, and eliminate spurious equilibria of complex-valued neural networks. One numerical example is provided to show the effectiveness and superiority of the presented results.
文摘Memristor is a newly found fourth circuit element for the next generation emerging nonvolatile memory technology. In this paper, design of new type of nonvolatile static random access memory cell is proposed by using a combination of memristor and complemented metal oxide semiconductor. Biolek memristor model and CMOS 180 nm technology are used to form a single cell. By introducing distinct binary logic to avoid safety margin is left for each binary logic output and enables better read/write data integrity. The total power consumption reduces from 0.407 mw (milli-watt) to 0.127 mw which is less than existing memristor based memory cell of the same CMOS technology. Read and write time is also significantly reduced. However, write time is higher than conventional 6T SRAM cell and can be reduced by increasing motion of electron in the memristor. The change of the memristor state is shown by applying piecewise linear input voltage.
文摘This paper presents the results of a 14-week attention strategy training of 174 college freshmen. It illustrates the promoting function of attention in second language vocabulary acquisition by raising students' expectations for new words, by increasing the frequency of exposure to them, by enhancing their perceptual salience, and by increasing the task demand for word study. The results show that enhancing attention in input could promote students' vocabulary acquisition and help them form vocabulary learning strategy suitable for their levels of proficiency.
文摘由于太阳能和天气变量的随机性和不稳定性,光伏发电具有很高的不确定性,准确的光伏发电功率预测对于光伏电站的短期调度和发电计划的运行至关重要。提出一种基于多模式分解、多通道输入、并联卷积神经网络(parallel convolutional neural network,PCNN)和双向长短期记忆网络(bi-directional long short term memory,BiLSTM)的组合预测方法,用于不同天气类型的超短期光伏发电功率预测。首先,由相关性分析算法确定辐照度和温度是对光伏发电贡献最大的2个环境变量,并根据环境因素与光伏功率波动特征的关联性将全年数据划分为4类;其次,使用完全集合经验模态分解、奇异谱分解和变分模态分解对辐照度、温度和光伏发电功率进行分解,以降低原始数据的复杂度和非平稳性,实现不同模式模态分量规律互补;最后,建立基于PCNN和BiLSTM的组合预测模型,使用PCNN提取不同的深度特征,并将PCNN输出的特征融合后输入到BiLSTM中,使用BiLSTM建立历史数据之间的时间特征关系,学习历史数据间的正、反向规律,在时空相关性分析的基础上得到最终光伏发电功率预测结果。实验结果表明,提出的组合预测方法在超短期光伏发电功率预测中具有较高的准确性和稳定性,并优于其他深度学习方法。