Training neural networks using Metropolis Monte Carlo and an adaptive variant

05/16/2022
by   Stephen Whitelam, et al.
0

We examine the zero-temperature Metropolis Monte Carlo algorithm as a tool for training a neural network by minimizing a loss function. We find that, as expected on theoretical grounds and shown empirically by other authors, Metropolis Monte Carlo can train a neural net with an accuracy comparable to that of gradient descent, if not necessarily as quickly. The Metropolis algorithm does not fail automatically when the number of parameters of a neural network is large. It can fail when a neural network's structure or neuron activations are strongly heterogenous, and we introduce an adaptive Monte Carlo algorithm, aMC, to overcome these limitations. The intrinsic stochasticity of the Monte Carlo method allows aMC to train neural networks in which the gradient is too small to allow training by gradient descent. We suggest that, as for molecular simulation, Monte Carlo methods offer a complement to gradient-based methods for training neural networks, allowing access to a distinct set of network architectures and principles.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
09/29/2017

User-friendly guarantees for the Langevin Monte Carlo with inaccurate gradient

In this paper, we revisit the recently established theoretical guarantee...
research
08/15/2020

Correspondence between neuroevolution and gradient descent

We show analytically that training a neural network by stochastic mutati...
research
02/17/2021

Multilevel Monte Carlo learning

In this work, we study the approximation of expected values of functiona...
research
10/27/2017

Automated Design using Neural Networks and Gradient Descent

We propose a novel method that makes use of deep neural networks and gra...
research
06/06/2019

Tensor Processing Units for Financial Monte Carlo

Monte Carlo methods are core to many routines in quantitative finance su...
research
10/10/2022

Rejecting noise in Baikal-GVD data with neural networks

Baikal-GVD is a large (∼ 1 km^3) underwater neutrino telescope installed...
research
06/19/2021

Rayleigh-Gauss-Newton optimization with enhanced sampling for variational Monte Carlo

Variational Monte Carlo (VMC) is an approach for computing ground-state ...

Please sign up or login with your details

Forgot password? Click here to reset