mlsanityData to 14 Oct 2026

Non-reversible Parallel Tempering for Deep Posterior Approximation

arXiv 2211.10837 ↗First posted 20 Nov 2022cs.LG stat.COcited by 7

Not on the map. Why?

Abstract (arXiv)

Parallel tempering (PT), also known as replica exchange, is the go-to workhorse for simulations of multi-modal distributions. The key to the success of PT is to adopt efficient swap schemes. The popular deterministic even-odd (DEO) scheme exploits the non-reversibility property and has successfully reduced the communication cost from O(P2) to O(P) given sufficiently many P chains. However, such an innovation largely disappears in big data due to the limited chains and few bias-corrected swaps. To handle this issue, we generalize the DEO scheme to promote non-reversibility and propose a few solutions to tackle the underlying bias caused by the geometric stopping time. Notably, in big data scenarios, we obtain an appealing communication cost O(Plog P) based on the optimal window size. In addition, we also adopt stochastic gradient descent (SGD) with large and constant learning rates as exploration kernels. Such a user-friendly nature enables us to conduct approximation tasks for complex posteriors without much tuning costs.

Key facts. Cited by 7 papers in the map (data through 14 Oct 2026). Builds on 26 of its 58 references that are in the map. First posted on arXiv 20 Nov 2022.

Builds on

26 of 58 references.

References · 26

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,926
  3. 2018Averaging Weights Leads to Wider Optima and Better Generalizationcited by 1,209
  4. 2019A Simple Baseline for Bayesian Uncertainty in Deep Learningcited by 560
  5. 2016Stein Variational Gradient Descent: A General Purpose Bayesian Inference Algorithmcited by 507
  6. 2016Entropy-SGD: Biasing Gradient Descent Into Wide Valleyscited by 359
  7. 2014Stochastic Gradient Hamiltonian Monte Carlocited by 343
  8. 2017Stochastic Gradient Descent as Approximate Bayesian Inferencecited by 300
  9. 2017Non-convex learning via Stochastic Gradient Langevin Dynamics: a nonasymptotic analysiscited by 243
  10. 2014Theoretical guarantees for approximate sampling from smooth and log-concave densitiescited by 208
  11. 2019Cyclical Stochastic Gradient MCMC for Bayesian Deep Learningcited by 203
  12. 2015Preconditioned Stochastic Gradient Langevin Dynamics for Deep Neural Networkscited by 121
14 more
  1. 2017Global Convergence of Langevin Dynamics Based Algorithms for Nonconvex Optimizationcited by 97
  2. 2016Statistical Inference for Model Parameters in Stochastic Gradient Descentcited by 71
  3. 2021Benign Overfitting of Constant-Stepsize SGD for Linear Regressioncited by 71
  4. 2020A statistical theory of cold posteriors in deep neural networkscited by 54
  5. 2015On Markov chain Monte Carlo methods for tall datacited by 50
  6. 2020Tight Nonparametric Convergence Rates for Stochastic Gradient Descent under the Noiseless Linear Modelcited by 39
  7. 2020Non-convex Learning via Replica Exchange Stochastic Gradient MCMCcited by 34
  8. 2013Austerity in MCMC Land: Cutting the Metropolis-Hastings Budgetcited by 24
  9. 2014Speeding Up MCMC by Efficient Data Subsamplingcited by 18
  10. 2020Accelerating Nonconvex Learning via Replica Exchange Langevin Diffusioncited by 16
  11. 2016An Efficient Minibatch Acceptance Test for Metropolis-Hastingscited by 11
  12. 2020Replica Exchange for Non-Convex Optimizationcited by 11
  13. 2020Accelerating Convergence of Replica Exchange Stochastic Gradient MCMC via Variance Reductioncited by 10
  14. 2022Interacting Contour Stochastic Gradient Langevin Dynamicscited by 9

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 2; 2025: 2; 2026: 3.

7 papers.

  1. 2024Diffusive Gibbs Samplingcited by 21
  2. 2024Constrained Exploration via Reflected Replica Exchange Stochastic Gradient Langevin Dynamicscited by 7
  3. 2025Active Learning with Bayesian Multi-Fidelity Laplace Neural Operators for Oscillatory Parametric PDEscited by 2
  4. 2025Enhancing Gradient-based Discrete Sampling via Parallel Temperingcited by 1
  5. 2026Position: The Time for Sampling Is Now! Charting a New Course for Bayesian Deep Learningcited by 1
  6. 2026Jeffreys Flow: Robust Boltzmann Generators for Rare Event Sampling via Parallel Tempering Distillationcited by 0
  7. 2026Explore Broadly, Reason Sharply: Push Small Models toward the Frontier via Samplingcited by 0