Pattern recognition minimizes entropy production in a neural network of electrical oscillators

Robert W. Hölzel, Katharina Krischer

Research output: Contribution to journalArticlepeer-review

13 Scopus citations

Abstract

We investigate the physical principle driving pattern recognition in a previously introduced Hopfield-like neural network circuit (Hölzel and Krischer, 2011 [13]). Effectively, this system is a network of Kuramoto oscillators with a coupling matrix defined by the Hebbian rule. We calculate the average entropy production 〈dS/dt〉 of all neurons in the network for an arbitrary network state and show that the obtained expression for 〈dS/dt〉 is a potential function for the dynamics of the network. Therefore, pattern recognition in a Hebbian network of Kuramoto oscillators is equivalent to the minimization of entropy production for the implementation at hand. Moreover, it is likely that all Hopfield-like networks implemented as open systems follow this mechanism.

Original languageEnglish
Pages (from-to)2766-2770
Number of pages5
JournalPhysics Letters, Section A: General, Atomic and Solid State Physics
Volume377
Issue number39
DOIs
StatePublished - 22 Nov 2013

Keywords

  • Global coupling
  • Oscillatory network
  • Time-dependent coupling
  • Weak coupling

Fingerprint

Dive into the research topics of 'Pattern recognition minimizes entropy production in a neural network of electrical oscillators'. Together they form a unique fingerprint.

Cite this