應用類神經網路模擬神經元細胞功能補償作用

The ANN Applied to Simulate the Functions Compensation Phenomena of Biological Neurons

朱明輝(Ming-Huei Chu) ； 曾國隆(Gow-Long Tzeng) ； 謝建新(Chien-Hsin Hsieh) ； 楊智勝(Zhi-Sheng Yang)

《機械技師學刊》 7卷4期 (2014/12) Pp. 8-14

https://doi.org/10.6208/JPME-2014-12-02

摘要

本研究探討基礎神經生物學，內容包含神經細胞、生物神經元及生物神經網路。簡述神經系統內神經元細胞構成及基礎功能及這些過程的整合作用。因為生物神經學的蓬勃發展使類神經網路（ANN）蓬勃發展，類神經網路係指模擬生物神經網路運作機制建立其數學模式，使得類神經網路可以具有學習及優化能力。本研究應用類神經網路中的倒傳遞網路（BPN）模擬生物神經元細胞老化後其功能逐漸被其他神經元細胞取代作用，以類神經網路模擬XOR邏輯閘功能，研究其學習過程收斂特性，並模擬單一生物神經元細胞老化失去功能後其他正常神經元細胞取代其失去功能之過程，並研究其學習過程收斂特性。

關鍵字

類神經網路 ； 倒傳遞網路 ； 細胞老化 ； 功能取代作用

並列摘要

This study introduces the biological neurology and artificial neural networks. The content of involves the neuron, nerve cell, the dissection, the physiology, the pharmacology, the pathology and the molecular biology. The development of biological neurology induces the attention of artificial neural networks (ANN), which is a mathematic model with the emulation capability for the functions of biological neural networks. The ANN can be applied intelligent commercial production by its learning capability. Because of the artificial biological neurology development, makes the neural network research to be possible to establish the mathematical model and the neural network research can lie in understood that the humanity nervous systems function. This study applied a BPN to emulate the XOR logical gate, the convergent processes investigated. The first neuron cell of the hidden layer loses function assumed, other neurons can take the lost functions of the first neuron, and the substitution processes are investigated by simulation.

並列關鍵字

Artificial neural networks ； Back propagation ； Aging cell ； Function substitution