Document Citation

Title:
An Information-Theoretic Neural Model Based On Concepts In Chinese Medicine.

Author(s):
Stefan Jaeger.

Institution(s):
1) National Library of Medicine, Bethesda, MD

Source:
Third International Workshop on Information Technology for Chinese Medicine, IEEE International Conference on Bioinformatics and Biomedicine (BIBM). October 2012:496-502.

Abstract:
This paper presents a linear information-theoretic model for neural signal processing using Chinese philosophical ideas, in particular Yin and Yang. The main goal is to provide a mathematical model that can explain neural activity in accordance with concepts in traditional Chinese medicine. As yet, the lack of such models has prevented a more formal explanation of the efficacy of traditional treatments related to the nervous system, such as acupuncture. According to the proposed model, a synapse performs a linear operation on its input. This operation, when plotted in polar coordinates, follows the shape of a Yin-Yang symbol. Synaptic learning becomes synonymous with adapting the size and rotation of the Yin-Yang symbol. The model distinguishes between the perceived input and the actual input. Both, perception and reality, coincide in the golden ratio. Learning thus becomes a process of aligning perception with reality, which is a novel learning concept.

Publication Type: CONFERENCE

More about this article:

Abstract | Full Text (PDF)