Neural network learning dynamics in a path integral framework
J. Balakrishnan
Abstract
A path-integral formalism is proposed for studying the dynamical evolution in time of patterns in an artificial neural network in the presence of noise. An effective cost function is constructed which determines the unique global minimum of the neural network system. The perturbative method discussed also provides a way for determining the storage capacity of the network.
Create a lesson
Related papers
Dynamically Reentrant Skyrmion Phase in Oscillating Magnetic Fields
Z. Demir Vatansever, E. Vatansever, A. Vasilopoulos et al.
Mutual linearity in semi-Markov networks
Julian B. Voits, Ulrich S. Schwarz
Steady states in asymmetric exclusion processes with reactions: Fluctuations and non-mean field behaviour
Rohn Chatterjee, Atri Goswami, Sudip Mukherjee
Marginal Operator Deformations and Pseudocriticality in Baxter--Wu Models
Hyun-Yong Lee, Naoki Kawashima, Nikolaos G. Fytas et al.
Kardar--Parisi--Zhang density hydrodynamics in a coherent XXZ--ASEP Lindbladian
Yansheng Liu, Daoguang Li, Kaixiang Lu et al.
Brownian gyrator model with stochastic resetting
Satya N. Majumdar, Gregory Schehr, Pascal Viot et al.