Interacting Contour Stochastic Gradient Langevin Dynamics
Machine Learning
2022-02-22 v1 Machine Learning
Abstract
We propose an interacting contour stochastic gradient Langevin dynamics (ICSGLD) sampler, an embarrassingly parallel multiple-chain contour stochastic gradient Langevin dynamics (CSGLD) sampler with efficient interactions. We show that ICSGLD can be theoretically more efficient than a single-chain CSGLD with an equivalent computational budget. We also present a novel random-field function, which facilitates the estimation of self-adapting parameters in big data and obtains free mode explorations. Empirically, we compare the proposed algorithm with popular benchmark methods for posterior sampling. The numerical results show a great potential of ICSGLD for large-scale uncertainty estimation tasks.
Cite
@article{arxiv.2202.09867,
title = {Interacting Contour Stochastic Gradient Langevin Dynamics},
author = {Wei Deng and Siqi Liang and Botao Hao and Guang Lin and Faming Liang},
journal= {arXiv preprint arXiv:2202.09867},
year = {2022}
}
Comments
ICLR 2022