Auto-Encoding Sequential Monte Carlo

By Tuan Anh Le, Maximilian Igl, Tom Rainforth, Tom Jin, Frank D. Wood. 2017

166 citations recorded by the source corpus.

We build on auto-encoding sequential Monte Carlo (AESMC): a method for model and proposal learning based on maximizing the lower bound to the log marginal likelihood in a broad family of structured probabilistic models. Our approach relies on the efficiency of sequential Monte Carlo (SMC) for performing inference in structured probabilistic models and the flexibility of deep neural networks to model complex conditional probability distributions. We develop additional theoretical insights and introduce a new training procedure which improves both model and proposal learning. We demonstrate that our approach provides a fast, easy-to-implement and scalable means for simultaneous model learning and proposal adaptation in deep generative models.

Subfields: Artificial Intelligence, Generative Models, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Generative Models

In Signal Processing

Citations in the corpus

Energy-Inspired Models: Learning with Sampler-Induced Distributions (2019, methodology): “Our second contribution is to show that EIMs provide a unifying conceptual framework to explain many advances in constructing tighter variational lower bounds for latent variable models ( e.g. , [9, 48, 51, 41, 68, 50, 63]).”
Deep Variational Reinforcement Learning for POMDPs (2018, methodology): “Le et al. (2018), Maddison et al. (2017) and Naesseth et al. (2018) introduce a new ELBO objective based on sequential Monte Carlo (SMC) (Doucet & Johansen, 2009) that allows faster learning in time series: ELBOSMC(θ, φ, o≤T ) = E [ T∑ t=1 log ( 1 K K∑ k=1 wkt )] , (6) where K is the number of…”
Importance Weighting and Variational Inference (2018, background): “Previous work [6, 2, 16, 12] investigated a similar connection between NIS and the importanceweighted ELBO.”
Reparameterization Gradient for Non-differentiable Models (2018, background): “One interesting future direction is to investigate the possibility of applying our ideas to recent variational objectives [24, 17, 19, 16, 25], which are based on tighter lower bounds of marginal likelihood than the standard ELBO.”
Filtering Variational Objectives (2017, background): “[38] additionally derive an expression for FIVO’s bias as the KL between the filter’s distribution and a certain target process.”
On Nesting Monte Carlo Estimators (2017, background): “As we show in Rainforth et al. (2018), these results further carry over to the reparameterized derivative estimates ∇ θ I N,M .”
The Thermodynamic Variational Objective (2019, methodology): “This objective is the same as the inference compilation objective (Le et al., 2018b).”
Variational Mixture-of-Experts Autoencoders for Multi-Modal Deep Generative Models (2019, methodology): “…maximising the standard ELBO indirectly minimises the KL between the variational and true posteriors, since Maximising the IWAE objective however, indirectly minimises the KL between implicit posteri-ors (Le et al., 2018) as where leading to higher-entropy estimates of the variational posterior.”
A General Method for Amortizing Variational Filtering (2018, methodology): “Finally, several recent works have used particle filtering in conjunction with amortized inference to provide a tighter lower bound on the log likelihood for sequential data [31, 33, 28].”
Importance Weighted Hierarchical Variational Inference (2019, background): “This bound has been shown (Domke and Sheldon, 2018) to be a tractable lower bound on ELBO for a variational distribution that has been obtained by self-normalizing importance sampling, or a special case of Sequential Monte Carlo (Maddison et al., 2017; Naesseth et al., 2017; Le et al., 2017).”
Divide and Couple: Using Monte Carlo Variational Objectives for Posterior Approximation (2019, background): “Recent work has investigated tighter objectives [6, 20, 18, 17, 23, 25], based on the following principle: Let R be an estimator of the likelihood—i.”
Tighter Variational Bounds are Not Necessarily Better (2018, methodology): “) is as per (2) and the form of the Kullback-Leibler ( KL ) is taken from (Le et al., 2018).”
State Space LSTM Models with Particle MCMC Inference (2017, methodology): “On a related note, sequential Monte Carlo methods have recently received attention in approximating the variational objective (Maddison et al., 2017; Le et al., 2017; Naesseth et al., 2017).”
Variational Sequential Monte Carlo (2017, methodology): “Finally, the log-marginal likelihood lower bound (6) was developed concurrently and independently by Maddison et al. [2017] and Le et al. [2017].”