mlsanityData to 7 Oct 2026

Variance reduction properties of the reparameterization trick

arXiv 1809.10330 ↗First posted 27 Sep 2018stat.ML cs.LG stat.COcited by 41

Not on the map. Why?

Abstract (arXiv)

The reparameterization trick is widely used in variational inference as it yields more accurate estimates of the gradient of the variational objective than alternative approaches such as the score function method. Although there is overwhelming empirical evidence in the literature showing its success, there is relatively little research exploring why the reparameterization trick is so effective. We explore this under the idealized assumptions that the variational approximation is a mean-field Gaussian density and that the log of the joint density of the model parameters and the data is a quadratic function that depends on the variational mean. From this, we show that the marginal variances of the reparameterization gradient estimator are smaller than those of the score function gradient estimator. We apply the result of our idealized analysis to real-world examples.

Key facts. Cited by 41 papers in the map (data through 7 Oct 2026). First posted on arXiv 27 Sep 2018.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 3; 2022: 7; 2023: 7; 2024: 11; 2025: 7; 2026: 6.

41 papers.

Sort:

By area:

  1. 2022Hierarchical Model-Based Imitation Learning for Planning in Autonomous Drivingcited by 43
  2. 2021Challenges and Opportunities in High-dimensional Variational Inferencecited by 35
  3. 2023Provable convergence guarantees for black-box variational inferencecited by 29
  4. 2023On the Convergence of Black-Box Variational Inferencecited by 22
  5. 2022PAC-Bayesian Offline Contextual Bandits With Guaranteescited by 19
  6. 2022Embedding Synthetic Off-Policy Experience for Autonomous Driving via Zero-Shot Curriculacited by 17
  7. 2021A unified view of likelihood ratio and reparameterization gradientscited by 12
  8. 2023Linear Convergence of Black-Box Variational Inference: Should We Stick the Landing?cited by 10
  9. 2021Model-free Policy Learning with Reward Gradientscited by 9
  10. 2024Bayesian Experimental Design via Contrastive Diffusionscited by 9
  11. 2023Practical and Matching Gradient Variance Bounds for Black-Box Variational Bayesian Inferencecited by 6
  12. 2024Understanding What Affects the Generalization Gap in Visual Reinforcement Learning: Theory and Empirical Evidencecited by 5
29 more
  1. 2024Provably Scalable Black-Box Variational Inference with Structured Variational Familiescited by 4
  2. 2025Hierarchical Multi-Agent Reinforcement Learning with Control Barrier Functions for Safety-Critical Autonomous Systemscited by 4
  3. 2023Fast post-process Bayesian inference with Variational Sparse Bayesian Quadraturecited by 3
  4. 2022Global convergence of optimized adaptive importance samplerscited by 2
  5. 2024Demystifying SGD with Doubly Stochastic Gradientscited by 2
  6. 2025Learning Hyperparameters via a Data-Emphasized Variational Objectivecited by 2
  7. 2025Improving Controller Generalization with Dimensionless Markov Decision Processescited by 2
  8. 2022Manifold Gaussian Variational Bayes on the Precision Matrixcited by 1
  9. 2023Parameter Estimation in DAGs from Incomplete Data via Optimal Transportcited by 1
  10. 2024Variance Control for Black Box Variational Inference Using The James-Stein Estimatorcited by 1
  11. 2024Learning the Regularization Strength for Deep Fine-Tuning via a Data-Emphasized Variational Objectivecited by 1
  12. 2025Nearly Dimension-Independent Convergence of Mean-Field Black-Box Variational Inferencecited by 1
  13. 2025Rao-Blackwellised Reparameterisation Gradientscited by 1
  14. 2026MixDPO: Modeling Preference Strength for Pluralistic Alignmentcited by 1
  15. 2026Revisiting Mixture Policies in Entropy-Regularized Actor-Criticcited by 1
  16. 2022Joint control variate for faster black-box variational inferencecited by 0
  17. 2022Approaching Globally Optimal Energy Efficiency in Interference Networks via Machine Learningcited by 0
  18. 2023Large-scale global optimization of ultra-high dimensional non-convex landscapes based on generative neural networkscited by 0
  19. 2024What to Do When Your Discrete Optimization Is the Size of a Neural Network?cited by 0
  20. 2024Diagonalisation SGD: Fast & Convergent SGD for Non-Differentiable Models via Reparameterisation and Smoothingcited by 0
  21. 2024Gradient Estimation and Variance Reduction in Stochastic and Deterministic Modelscited by 0
  22. 2024You Only Accept Samples Once: Fast, Self-Correcting Stochastic Variational Inferencecited by 0
  23. 2024Generalizing soft actor-critic algorithms to discrete action spacescited by 0
  24. 2025Classifier Weighted Mixture modelscited by 0
  25. 2025Distributions as Actions: A Unified Framework for Diverse Action Spacescited by 0
  26. 2026Stochastic Gradient Variational Inference with Price's Gradient Estimator from Bures-Wasserstein to Parameter Spacecited by 0
  27. 2026Does "Do Differentiable Simulators Give Better Policy Gradients?'' Give Better Policy Gradients?cited by 0
  28. 2026Deconstructing Actor-Critic: A Large-scale Empirical Study of Design Components for Practitionerscited by 0
  29. 2026Accelerating Evolutionary Strategy via Rao-Blackwellizing Realization of Uncertain Inputcited by 0