AAAI Publications, 2017 AAAI Spring Symposium Series

Font Size: 
Feynman Machine: A Novel Neural Architecture for Cortical and Machine Intelligence
Eric Laukien, Richard Crowder, Fergal Byrne

Last modified: 2017-03-20

Abstract


Developments in the study of Nonlinear Dynamical Systems (NDS's) over the past thirty years have allowed access to new understandings of natural and artificial phenomena, yet much of this work remains unknown to the wider scientific community. In particular, the fields of Computational Neuroscience and Machine Learning rely heavily for their theoretical basis on ideas from 19th century Statistical Physics, Linear Algebra, and Statistics, which neglect or average out the important information content of time series signals generated between and within NDS's. In contrast, the Feynman Machine, our model of cortical and machine intelligence, is designed specifically to exploit the computational power of coupled, communicating NDS's. Recent empirical evidence of causal coupling in primate neocortex corresponds closely with our model. A high-performance software implementation has been developed, allowing us to examine the computational properties of this novel Machine Learning framework.

Keywords


Neural Computation; Machine Learning; Dynamical Systems; Artificial Intelligence

Full Text: PDF