1.School of Computer Science and Engineering, Northwestern Polytechnical University, Xi’an 710072, China 2.School of Electronics and Information, Northwestern Polytechnical University, Xi’an 710072, China
Abstract:The behavior of transition from short-term memory (STM) to long-term memory (LTM) has been observed and reported in the experimental studies of memristors fabricated by different materials. This kind of memristor in this paper is named STM→LTM memristor. In some of these experimental researches, the learning-experience behavior observed in the " learning-forgetting-relearning” experiment is also reported. When the memristor is restimulated by pulses after forgetting the STM, its memory will quickly return to the highest state that has been reached before the forgetting period, and the memory recovery during the relearning period is obviously faster than the memory formation in the first learning process. In this paper, the behavior of the existing STM→LTM memristor model in the " learning-forgetting-relearning” experiment is further discussed. If wmax, the upper bound of the memory level, is a constant with a value of 1, the STM→LTM memristor model exhibits no learning-experience behavior, and this model shows a faster relearning behavior in the " learning-forgetting-relearning” experiment. The relearning process is faster because the memory forgetting during pulse-to-pulse interval in the relearning process is slower than that in the first learning process. In the STM→LTM memristor model with learning-experience behavior, wmax is redesigned as a state variable in [0,1], and its value will be influenced by the applied voltage. The memory formation in the first learning process is relatively slow because wmax limits the memory formation speed when the pulse is applied. After the forgetting process, the limitation of wmax on the pulse-induced memory formation is less obvious, so the memory of the device increases at a faster speed during the memory recovery of the relearning process. In this case, the forgetting speed still becomes slower after each pulse has been applied. If the pulse-induced wmax increase is so fast that wmax will quickly increase to its upper bound after a few pulses have been applied in the first learning process, and the learning-experience behavior is similar to the faster relearning behavior when wmax = 1. In most of experimental research papers about the STM→LTM memristor, the change of the memristance can be explained by the formation and annihilation of the conductive channel between two electrodes of a memristor. During a certain period of time, the ions (or vacancies), which can be used to form the conductive channel, are only those that are around the conductive channel, which indicates that there should be an upper bound for the size of the conductive channel within this time period. The area in which ions (or vacancies) can be used to form the conductive channel is called the surrounding area of the conductive channel. In the model, wmax can be understood as the size of the conductive channel’s surrounding area, and it describes the upper bound of the width of the conductive channel. Keywords:memristor/ learning-experience behavior/ short-term memory/ long-term memory
a) “记忆遗忘量”是根据各个文献中所报道的“学习-遗忘-再学习”实验的遗忘曲线, 对于记忆的遗忘量占初次学习所形成的记忆总量的比例给出估计, 各个数据前的“~”表示该数据是根据实验曲线所得的估计值. b) 在文献[27—30]所报道的实验中经历了多次“遗忘”与“再学习”过程, 表中对这几篇文献依次给出每次遗忘过程的遗忘比例, 以及每次再学习阶段的记忆恢复所经历的脉冲数量.
表1文献[21—30]中所报道的“学习-遗忘-再学习”实验数据 Table1.Experimental data of “learning-forgetting-relearning” experiment in References [21—30].
模型中的状态变量w, wmin, wmax在脉冲电压作用时的最大增速分别由各自状态方程中的参数τw+, τmin+, τmax+所描述. 图5—图7给出参数τw+, τmin+, τmax+大小的改变对于“学习-遗忘-再学习”实验中状态变量w, wmin, wmax的变化情况的影响, 仿真中除了参数τw+, τmin+, τmax+以外, τw0_min = 0.1 s, kτ+ = 50, τmax 0 = 105 s, 其他模型参数以及所施加的脉冲参数与图4仿真中的参数取值相同. “学习-遗忘-再学习”实验中忆阻器依次经历: 30个脉冲的初次学习、100 s的遗忘过程、30个脉冲的再次学习. 仿真初始时刻的各个状态变量均处于其下限. 在初次学习阶段w所达到的最大值在图中用虚线标出. 图 5τw+ = 0.06 s, τmax+ = 0.05 s, τmin+取不同值时的“学习-遗忘-再学习”实验仿真结果 (a) τmin+ = 0.5 s; (b) τmin+ = 0.2 s; (c) τmin+ = 0.1 s. 虚线标出w在初次学习阶段所达到的最大值 Figure5. Simulations of “learning-forgetting-relearning” process when τw+ = 0.06 s, τmax+ = 0.05 s, and τmin+ takes different values: (a) τmin+ = 0.5 s; (b) τmin+ = 0.2 s; (c) τmin+ = 0.1 s. The maximum value of w during the first learning period is marked by a dashed line.
图 7τw+ = 0.15 s, τmin+ = 0.1 s, τmax+取不同值时的“学习-遗忘-再学习”实验仿真结果 (a) τmax+ = 0.1 s; (b) τmax+ = 0.05 s; (c) τmax+ = 0.003 s; (d) 当wmax为常数1, 模型其他参数与(a)?(c)相同时的“学习-遗忘-再学习”实验仿真结果. 虚线标出w在初次学习阶段所达到的最大值 Figure7. Simulations of “learning-forgetting-relearning” process when τw+ = 0.15 s, τmin+ = 0.1 s, and τmax+ takes different values: (a) τmax+ = 0.1 s; (b) τmax+ = 0.05 s; (c) τmax+ = 0.003 s; (d) simulation of “learning-forgetting-relearning” process when wmax = 1, and other parameters are the same as those used in (a)?(c). The maximum value of w during the first learning period is marked by a dashed line.
参数τmin+描述了wmin在脉冲激励作用时的最大增速: 当其他参数一定时, τmin+越小, wmin在脉冲激励作用时的增长越快. 图5给出τw+ = 0.06 s, τmax+ = 0.05 s, τmin+取不同值时的“学习-遗忘-再学习”实验仿真结果. 在初次学习阶段, τmin+取值的不同会对wmin的增速有明显影响, 而对于w和wmax的增长速度影响并不明显. 在之后的遗忘阶段, wmin和wmax基本保持不变, w将逐渐递减并收敛于wmin. 由于τmin+取不同值时在初次学习阶段wmin的增量不同, 因此在遗忘阶段w的收敛值以及在再学习阶段的记忆恢复所需的脉冲数量也有相应的区别: τmin+取值分别为0.5, 0.2和0.1 s时, 记忆恢复所经历的脉冲数量分别为16, 14和12个. 参数τw+描述了w在脉冲作用时的最大增长速度: 当其他参数一定时, τw+越小, w在脉冲激励作用时的增长越快. 图6给出τmin+ = 0.2 s, τmax+ = 0.1 s, τw+取不同值时的“学习-遗忘-再学习”实验仿真结果. 在初次学习阶段, τw+越小, 则w的增长过程越接近于其上限wmax的增长过程; τw+的大小对于wmax的变化速度并无明显影响; w限制了wmin的变化上限, 当w增长速度变快时, wmin的增速也会有所增长, 因此τw+的减小也会使得w在遗忘过程的收敛值增大. 在再次学习阶段, 当τw+取值分别为0.15和0.06 s时, 如图6(a)和图6(b)所示, 完成记忆恢复需分别经历12和11次脉冲激励, 记忆恢复前后脉冲作用下的记忆增速无明显变化; 当τw+ = 0.01 s时, 如图6(c)所示, 仅经历5次脉冲激励后就可使得w恢复在初次学习中所达到的最大值, 在之后的脉冲激励作用下w将跟随wmax逐渐增长, 脉冲作用下的记忆增速在记忆恢复之后明显比记忆恢复阶段更小. 文献[21,30]中报道了再学习阶段记忆恢复后脉冲继续作用下的记忆增长情况: 文献[21]所报道的实验现象中, 再学习阶段记忆恢复速度比初次学习的记忆增速快得多, 在记忆恢复后记忆的增速会有明显的减慢, 减慢后的记忆增速与初次学习的记忆增速相似; 文献[30]中的忆阻器在再学习阶段的记忆恢复速度略快于初次学习的记忆增速, 记忆恢复后脉冲继续作用下的记忆增速并无明显变化. 由图6所示仿真过程可知, 这两种情况的实验现象可通过调整参数τw+的大小来描述. 图 6τmin+ = 0.2 s, τmax+ = 0.1 s, τw+取不同值时的“学习-遗忘-再学习”实验仿真结果 (a) τw+ = 0.15 s; (b) τw+ = 0.06 s; (c) τw+ = 0.01 s. 虚线标出w在初次学习阶段所达到的最大值 Figure6. Simulations of “learning-forgetting-relearning” process when τmin+ = 0.2 s, τmax+ = 0.1 s, and τw+ takes different values: (a) τw+ = 0.15 s; (b) τw+ = 0.06 s; (c) τw+ = 0.01 s. The maximum value of w during the first learning period is marked by a dashed line.