On strict convergence of stochastic gradients

2016/10/11 by Benaim, Michel
#Dynamical Systems (math.DS) #FOS: Mathematics #Probability (math.PR)

paper · doi:10.48550/arxiv.1610.03278

Abstract

We discuss conditions ensuring the (strict) convergence of stochastic gradient algorithms.

Related

• On Almost Sure Convergence Rates of Stochastic Gradient Methods
• A Stochastic Gradient Descent Theorem and the Back-Propagation Algorithm
• New Convergence Aspects of Stochastic Gradient Algorithms
• On the linear convergence of the stochastic gradient method with constant step-size
• Handbook of Convergence Theorems for (Stochastic) Gradient Methods