no code implementations • 8 Nov 2023 • Neil K. Chada, Benedict Leimkuhler, Daniel Paulin, Peter A. Whalley
We exhibit similar bounds using both approximate and stochastic gradients, and our method's computational cost is shown to scale logarithmically with the size of the dataset.
1 code implementation • 20 Jun 2021 • Tiffany Vlaar, Benedict Leimkuhler
We also discuss splitting choices for the neural network parameters which could enhance generalization performance when neural networks are trained from scratch.
1 code implementation • 20 Jun 2021 • Benedict Leimkuhler, Tiffany Vlaar, Timothée Pouchon, Amos Storkey
We employ constraints to control the parameter space of deep neural networks throughout training.
no code implementations • 17 Jun 2020 • Benedict Leimkuhler, Timothée Pouchon, Tiffany Vlaar, Amos Storkey
We propose a method for efficiently incorporating constraints into a stochastic gradient Langevin framework for the training of deep neural networks.
1 code implementation • 30 Aug 2019 • Benedict Leimkuhler, Charles Matthews, Tiffany Vlaar
We describe easy-to-implement hybrid partitioned numerical algorithms, based on discretized stochastic differential equations, which are adapted to feed-forward neural networks, including a multi-layer Langevin algorithm, AdLaLa (combining the adaptive Langevin and Langevin algorithms) and LOL (combining Langevin and Overdamped Langevin); we examine the convergence of these methods using numerical studies and compare their performance among themselves and in relation to standard alternatives such as stochastic gradient descent and ADAM.
no code implementations • 20 Mar 2019 • Frederik Heber, Zofia Trstanova, Benedict Leimkuhler
With the advent of GPU-assisted hardware and maturing high-efficiency software platforms such as TensorFlow and PyTorch, Bayesian posterior sampling for neural networks becomes plausible.
1 code implementation • 13 Jul 2016 • Charles Matthews, Jonathan Weare, Benedict Leimkuhler
We describe parallel Markov chain Monte Carlo methods that propagate a collective ensemble of paths, with local covariance information calculated from neighboring replicas.
no code implementations • NeurIPS 2015 • Xiaocheng Shang, Zhanxing Zhu, Benedict Leimkuhler, Amos J. Storkey
Monte Carlo sampling for Bayesian posterior inference is a common approach used in machine learning.
1 code implementation • 24 Mar 2012 • Benedict Leimkuhler, Charles Matthews
We then compare Langevin dynamics integrators in terms of their invariant distributions and demonstrate a superconvergence property (4th order accuracy where only 2nd order would be expected) of one method in the high friction limit; this method, moreover, can be reduced to a simple modification of the Euler-Maruyama method for Brownian dynamics involving a non-Markovian (coloured noise) random process.
Numerical Analysis Statistical Mechanics Chemical Physics Computational Physics