2014/04/23 by Kenric P. Nelson, Madalina Barbu, Brian J. Scannell
Computer Science · Mathematics · Physics and Astronomy · #Artificial neural network #Gaussian Processes and Bayesian Inference #Generalization #Graphical model #Markov process #Neural Networks and Applications #Node (physics) #Nonlinear system #Probabilistic logic #Random field #Random variable #Statistical Mechanics and Entropy #Statistical model #cs.IT #cs.LG #cs.NE #math.IT
paper · pdf · doi:10.1117/12.2050759
Submitted for presentation at the Machine Intelligence and Bio-inspired Computation: Theory and Applications Conference, SPIE Sensing Technology and Applications, Baltimore, MD, May 8, 2014
arxiv created 2014/04/23 · openalex publication_date 2014/05/22 · arxiv updated 2015/06/19 · openalex created_date 2016/06/24 · openalex updated_date 2026/08/05
Neural network design has utilized flexible nonlinear processes which can mimic biological systems, but has suffered from a lack of traceability in the resulting network. Graphical probabilistic models ground network design in probabilistic reasoning, but the restrictions reduce the expressive capability of each node making network designs complex. The ability to model coupled random variables using the calculus of nonextensive statistical mechanics provides a neural node design incorporating nonlinear coupling between input states while maintaining the rigor of probabilistic reasoning. A generalization of Bayes rule using the coupled product enables a single node to model correlation between hundreds of random variables. A coupled Markov random field is designed for the inferencing and classification of UCI’s MLR ‘Multiple Features Data Set’ such that thousands of linear correlation parameters can be replaced with a single coupling parameter with just a (3%, 4%) reduction in (classification, inference) performance.