为了提升机动目标的航迹预测精准度,提出了一种粒子群(Particle Swarm Optimization,PSO)算法与长短周期记忆单元(Long Short Term Memory,LSTM)神经网络相结合的PSO-LSTM目标航迹预测模型。针对LSTM神经网络中的超参数难以人工最优化...为了提升机动目标的航迹预测精准度,提出了一种粒子群(Particle Swarm Optimization,PSO)算法与长短周期记忆单元(Long Short Term Memory,LSTM)神经网络相结合的PSO-LSTM目标航迹预测模型。针对LSTM神经网络中的超参数难以人工最优化的问题,通过PSO算法进行自动调参,对诸如LSTM隐藏层规模、学习率、训练集批次规模等参数进行优化。使用PSO-LSTM航迹预测模型在真实航迹数据集上进行预测,实验结果表明,PSO-LSTM航迹预测模型在MAE、RMSE等指标上优于传统模型,有着较好的准确性与稳定性。展开更多
To improve students＇ reading skill through VMQ4R-EIM and to reflect the English instructional process implemented by the teacher in the secondary school, and finally the teacher can introspect his English instruction...To improve students＇ reading skill through VMQ4R-EIM and to reflect the English instructional process implemented by the teacher in the secondary school, and finally the teacher can introspect his English instructional activities, there are 40 students as the research subjects of this CAR （classroom action research） and it used spriral method from Kemmis and McTaggart with four stages （plan, action, observation, reflection）. The research result showed that there was significant reading skill improvement to secondary school students. In the first cycle, the student＇s reading skill improved to 62% and the observation result reached to 65%. In the second cycle, the student reading skill improved to 97% and the observation result reached to 95%. The VMQ4R-EIM can improve secondary school students＇ reading skill. The research implication is the VMQ4R-EIM that is one way to be implemented by the English teacher in order to improve and make the students motivated and active in attending the instructional process展开更多
Emerging memristors can be used as artificial synapses for emulating memory and computational functions.In this work,inspired by the memristive properties of tantalum dioxide,we designed a memristor with a structure o...Emerging memristors can be used as artificial synapses for emulating memory and computational functions.In this work,inspired by the memristive properties of tantalum dioxide,we designed a memristor with a structure of TiN/Ta2O5-x/HfxZr1-xO2(x=0.5)/Pt(TTHZOP).The device conductance can be continuously tuned by adjusting the voltage pulse parameters(i.e.,amplitude,width,and number)of voltage sweeps.Furthermore,for both of negative and positive parts,the current-voltage curves of the sweep cycle appear to better adjust the gradual distribution in successive twenty cycles.According to the fine fitting results of twenty positive and negative current-voltage(I-V)curves,the probability of an electron jumping over an energy barrier and the width of the energy barrier were analyzed in detail.It is found that the electron tunneling mechanism at the interface is responsible for gradual conduction change under successive external electrical stimulation consisting of both bulk and interface effects.The proposed TTHZOP memristor is a promising candidate in potential applications that mimic artificial biosynaptic adaptation and analog brain computation.展开更多
文摘为了提升机动目标的航迹预测精准度,提出了一种粒子群(Particle Swarm Optimization,PSO)算法与长短周期记忆单元(Long Short Term Memory,LSTM)神经网络相结合的PSO-LSTM目标航迹预测模型。针对LSTM神经网络中的超参数难以人工最优化的问题,通过PSO算法进行自动调参,对诸如LSTM隐藏层规模、学习率、训练集批次规模等参数进行优化。使用PSO-LSTM航迹预测模型在真实航迹数据集上进行预测,实验结果表明,PSO-LSTM航迹预测模型在MAE、RMSE等指标上优于传统模型,有着较好的准确性与稳定性。
文摘To improve students＇ reading skill through VMQ4R-EIM and to reflect the English instructional process implemented by the teacher in the secondary school, and finally the teacher can introspect his English instructional activities, there are 40 students as the research subjects of this CAR （classroom action research） and it used spriral method from Kemmis and McTaggart with four stages （plan, action, observation, reflection）. The research result showed that there was significant reading skill improvement to secondary school students. In the first cycle, the student＇s reading skill improved to 62% and the observation result reached to 65%. In the second cycle, the student reading skill improved to 97% and the observation result reached to 95%. The VMQ4R-EIM can improve secondary school students＇ reading skill. The research implication is the VMQ4R-EIM that is one way to be implemented by the English teacher in order to improve and make the students motivated and active in attending the instructional process
基金supported by the National Natural Science Foundation of China(61674050 and 61874158)the Outstanding Youth Funding of Hebei University(A2018201231)+2 种基金the Support Program for the Top Young Talents of Hebei Province(70280011807)the Hundred Persons Plan of Hebei Province(E2018050004 and E2018050003)the Supporting Plan for 100 Excellent Innovative Talents in Colleges and Universities of Hebei Province(SLRC2019018)。
文摘Emerging memristors can be used as artificial synapses for emulating memory and computational functions.In this work,inspired by the memristive properties of tantalum dioxide,we designed a memristor with a structure of TiN/Ta2O5-x/HfxZr1-xO2(x=0.5)/Pt(TTHZOP).The device conductance can be continuously tuned by adjusting the voltage pulse parameters(i.e.,amplitude,width,and number)of voltage sweeps.Furthermore,for both of negative and positive parts,the current-voltage curves of the sweep cycle appear to better adjust the gradual distribution in successive twenty cycles.According to the fine fitting results of twenty positive and negative current-voltage(I-V)curves,the probability of an electron jumping over an energy barrier and the width of the energy barrier were analyzed in detail.It is found that the electron tunneling mechanism at the interface is responsible for gradual conduction change under successive external electrical stimulation consisting of both bulk and interface effects.The proposed TTHZOP memristor is a promising candidate in potential applications that mimic artificial biosynaptic adaptation and analog brain computation.
