# Learning Deep Generative Models with Annealed Importance Sampling

arXiv: https://arxiv.org/abs/1906.04904 · HTML: https://mlsanity.com/paper/1906.04904/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Jun 2019.

## Built on by (most cited first, 8 of 8)

- [MCMC Variational Inference via Uncorrected Hamiltonian Annealing](https://mlsanity.com/paper/2107.04150/) (2021)
- [Monte Carlo Variational Auto-Encoders](https://mlsanity.com/paper/2106.15921/) (2021)
- [Efficient Training of Energy-Based Models Using Jarzynski Equality](https://mlsanity.com/paper/2305.19414/) (2023)
- [Surrogate Likelihoods for Variational Annealed Importance Sampling](https://mlsanity.com/paper/2112.12194/) (2021)
- [Bootstrap Your Flow](https://mlsanity.com/paper/2111.11510/) (2021)
- [Ensemble-Based Annealed Importance Sampling](https://mlsanity.com/paper/2401.15645/) (2024)
- [NEO: Non Equilibrium Sampling on the Orbit of a Deterministic Transform](https://mlsanity.com/paper/2103.10943/) (2021)
- [Jarzynski Reweighting and Sampling Dynamics for Training Energy-Based Models: Theoretical Analysis of Different Transition Kernels](https://mlsanity.com/paper/2506.07843/) (2025)

## Abstract (arXiv)

Variational inference (VI) and Markov chain Monte Carlo (MCMC) are two main approximate approaches for learning deep generative models by maximizing marginal likelihood. In this paper, we propose using annealed importance sampling for learning deep generative models. Our proposed approach bridges VI with MCMC. It generalizes VI methods such as variational auto-encoders and importance weighted auto-encoders (IWAE) and the MCMC method proposed in (Hoffman, 2017). It also provides insights into why running multiple short MCMC chains can help learning deep generative models. Through experiments, we show that our approach yields better density models than IWAE and can effectively trade computation for model accuracy without increasing memory cost.