Japan (JSPS) Postdoctoral Program: Dynamics in Chaotic Neural Networks

该奖项将提供补充支持，使哈佛医学院的Elizabeth kerman博士能够与日本东京电机大学的K. Aihara博士进行为期12个月的合作研究。他们将通过分析由具有混沌动力学的神经元组成的人工神经网络的特点，探讨混沌在神经信息处理中的意义。确定性混沌的非线性现象在反馈系统、光学、生态学和生物系统等领域普遍存在。这个项目将尝试展示混沌波动在网络和神经计算应用中的影响。例如，由于混沌神经元可以自发地产生宏观波动，它们可以在优化、学习和自组织问题中提供一种摆脱不希望的局部最小值的方法。爱原博士在混沌神经元数学建模方面的背景将补充克勒曼博士在昼夜节律领域的研究经验。

This award will provide supplementary support to enable Dr. Elizabeth Klerman of the Harvard Medical School to conduct collaborative research for 12 months with Dr. K. Aihara of Tokyo Denki University, Japan. They will investigate the significance of chaos in neural information processing by analyzing the characteristics of artificial neural networks composed of neurons with chaotic dynamics. The nonlinear phenomena of deterministic chaos is now recognized as ubiquitous in fields such as feedback systems, optics, ecology and biological systems. This project will attempt to demonstrate the effects of chaotic fluctuations in applications of network and neurocomputing. For example, since chaotic neurons can produce macroscopic fluctuations spontaneously, they may provide a means of escaping from undesirable local minima in problems of optimiza- tion, learning and self organization. Dr. Aihara's background in mathematical modeling of chaotic neurons will complement Dr. Klerman's research experience in the area of circadian rhythms.