中国科学院自动化研究所机构知识库
Advanced  
CASIA OpenIR  > 类脑智能研究中心  > 会议论文
题名: A Spiking Neural Network Based Autonomous Reinforcement Learning Model and Its Application in Decision Making
作者: Wang Guixiang(王桂香)1; Yi Zeng1, 2; Bo Xu1, 2
出版日期: 2016-11
会议名称: BICS 2016: The 8th International Conference on Brain Inspired Cognitive Systems
会议日期: 2016年11月27-30日
会议地点: Beijing, China
关键词: Spiking neural network ; Hodgkin-Huxley ; Basal Ganglia ; motion perception
学科分类: 交叉与边缘领域的力学
通讯作者: Yi Zeng
英文摘要:

In this paper, we propose an autonomous spiking neural network model for decision making. The model is an expansion of the basal ganglia circuitry with automatic environment perception, which constructs environmental states automatically from image inputs. The work in this paper has the following contributions: (1) In our model, the simplified Hodgkin-Huxley computing model is developed to achieve calculation efficiency closed to the LIF model and is used to obtain and test the ionic level properties in cognition. (2) A spike based motion perception mechanism is proposed to extract key elements for learning process from raw pixels without large amount of training. We apply our model in the “flappy bird” game and it play well after dozens of trainings. The model gets similar learning performance with human at the start of training. Besides, our model simulates cognitive defects when blocking some of sodium or potassium ion channels in the Hodgkin-Huxley model and this is an exploration of cognition deep into ionic level


收录类别: CPCI-T
会议录: Conferences on the 8th International Conference on Brain-inspired Cognitive System
内容类型: 会议论文
URI标识: http://ir.ia.ac.cn/handle/173211/12619
Appears in Collections:类脑智能研究中心_会议论文

Files in This Item: