Deformed Statistics Free Energy Model for Source Separation using Unsupervised Learning

TL;DR

This paper introduces a novel unsupervised learning model for source separation based on a deformed statistics free energy approach utilizing Tsallis' entropy and q-deformed calculus, demonstrating its effectiveness through numerical examples.

Contribution

It presents a new variational principle for source separation using Tsallis' entropy and integrates it with Hopfield-like learning rules via q-deformed calculus.

Findings

Effective source separation demonstrated in numerical examples

Novel integration of Tsallis' entropy with Hopfield-like learning rules

Shows potential for improved unsupervised learning models

Abstract

A generalized-statistics variational principle for source separation is formulated by recourse to Tsallis' entropy subjected to the additive duality and employing constraints described by normal averages. The variational principle is amalgamated with Hopfield-like learning rules resulting in an unsupervised learning model. The update rules are formulated with the aid of q-deformed calculus. Numerical examples exemplify the efficacy of this model.