In addition to the shared weights of synaptic connections, PNN includes weights of synaptic ranges for Forward propagation and Back propagation[15,16,19-24]. PNN considers synaptic strength balance in dynamic of phagocytosing of synapses and static of constant sum of synapses length [15],the lead behavior of the school of fish is well embodied in our PNN. Synapse formation will inhibit dendrites generation to a certain extent in experiments, by simulations synapse formation will inhibit the function of dendrites [16]. Closing the critical period will cause neurological disorder in experiments, but worse results in PNN simulations [19]. The memory persistence gradient information of backward circuit similar to the Enforcing Resilience in a Spring Boot. The relatively good and inferior gradient information in synapse formation of backward circuit like the folds of the brain. Considering both negative and positive memories persistence help activate synapse length changes with iterations better than only considering positive memory. So using memory of fear learning and improving of synaptic activity to observe obviously [20]. Memory persistence factor also inhibit local synaptic accumulation. And refers PNN can also introduce the relatively good and inferior solution to update the velocity of particle in PSO. Astrocytic phagocytosis will avoid the local accumulation of synapses by simulation (Lack of astrocytic phagocytosis causes excitatory synapses and functionally impaired synapses accumulate in experiments and lead to destruction of cognition, but local longer synapses and worse results in PNN simulations) [21]. It gives relationship of human intelligence and cortical thickness, individual differences in brain[22].PNN also considered the memory engram cells that strengthened Synaptic strength[23]. The simple PNN which only has the synaptic phagocytosis.
翻译:除了共同的突触连接权重外, PNN 还包括前方传播和后方传播的突变范围[ 15, 16, 19-24] 。 PNN 考虑突触突变动态中的合成强度平衡以及恒定突触长度[15] 的静态和恒定总和[15] 。鱼群中的铅行为在我们的 PNN 中得到了很好的体现。 Synapse 形成将在某种程度上抑制在实验中产生畸形。 通过模拟突触形成,将抑制 dendrates [16] 。 关键时期的结束将导致实验中的神经紊乱,但PNNNN的模拟结果则更糟糕。 后向电路的内向梯度信息与Spray Boot相类似。 脉冲的后向电流形成相对良好和低劣的梯度信息, 考虑到负和正向的记忆,它只会有助于通过直径直的记忆来激活振动的神经变变变变变。 因此, 利用对恐惧的记忆, 精度的内径直变变变的内径变,, 也意味着的内径变变的内径变变的内积 。