arXiv research

Square-root natural-gradient improves variational inference convergence.

Variational inference approximates the posterior distribution of a probabilistic model with a parameterized density by maximizing a lower bound for the model evidence. Modern solutions fit a flexible approximation with stochastic gradient descent, using Monte Carlo approximation for the gradients. This enables variatio…

Paper analyzes inclusive KL inference using Wasserstein gradient flows.

This paper shows equivalence between SVGD and BBVI using kernel gradient flows.

This paper bridges variational inference and Wasserstein gradient flows.

NVGD uses neural networks to infer distributions without kernel choices.

A new method for efficient inference in probabilistic programs with mixed support.

Unified framework for statistical inference in gradient boosting regression.

Bayesian inference plays an important role in advancing machine learning, but faces computational challenges when applied to complex models such as deep neural networks. Variational inference circumvents these challenges by formulating Bayesian inference as an optimization problem and solving it using gradient-based op…

A new ParVI framework improves particle-based variational inference methods.

Variational inference transforms posterior inference into parametric optimization thereby enabling the use of latent variable models where otherwise impractical. However, variational inference can be finicky when different variational parameters control variables that are strongly correlated under the model. Traditiona…

This work improves variational inference by reducing gradient variance.

This work proposes a new method for variational inference using Wasserstein gradient descent.

Variational inference has experienced a recent surge in popularity owing to stochastic approaches, which have yielded practical tools for a wide range of model classes. A key benefit is that stochastic variational inference obviates the tedious process of deriving analytical expressions for closed-form variable updates…

New method reduces inference variance for faster optimization.

Unified framework analyzes privacy risks from gradients in distributed learning.

Improves understanding of stochastic NGVI convergence rates.

Inspired by the seminal work on Stein Variational Inference and Stein Variational Policy Gradient, we derived a method to generate samples from the posterior variational parameter distribution by \textit{explicitly} minimizing the KL divergence to match the target distribution in an amortize fashion. Consequently, we a…

Enhances optimization and sampling methods using ensemble-based gradient inference.

GeoPhy uses geometric gradients to efficiently infer phylogenetic trees from molecular data.

The paper develops statistical inference for gradient flows in optimization.

This work proposes new methods for variational inference using gradient flows on Gaussian measures.

This work proposes using zero-variance control variates to reduce variance in pathwise gradient estimators for variational inference.

Natural-gradient methods enable fast and simple algorithms for variational inference, but due to computational difficulties, their use is mostly limited to \emph{minimal} exponential-family (EF) approximations. In this paper, we extend their application to estimate \emph{structured} approximations such as mixtures of E…

VarGrad reduces variance in ELBO gradient estimation for variational inference.

U-statistics improve gradient estimation in importance-weighted variational inference.

Quantized Variational Inference improves ELBO optimization with fast convergence.

SING improves state inference in latent SDE models for better drift function estimation.

A method for efficient approximate inference on discrete distributions.

New framework improves variational inference with Markov chain methods.

Stein variational gradient descent improves inference in Gaussian process models.

New guarantees for black-box variational inference methods.

New bounds show BBVI's gradient variance matches SGD conditions, improving parameterization efficiency.

Over the past decade there has been considerable interest in spectral algorithms for learning Predictive State Representations (PSRs). Spectral algorithms have appealing theoretical guarantees; however, the resulting models do not always perform well on inference tasks in practice. One reason for this behavior is the m…

We propose a technique for increasing the efficiency of gradient-based inference and learning in Bayesian networks with multiple layers of continuous latent vari- ables. We show that, in many cases, it is possible to express such models in an auxiliary form, where continuous latent variables are conditionally determini…

DGPs with variational inference suffer from SNR issues that degrade gradient estimates, leading to unreliable training.

New method controls gradient error for sparse MRFs.

BOED improves SBI by optimizing experimental designs and inference functions.

Improves learning of spectral mixture kernels with approximate Bayesian inference.

We present a novel approximate inference method for diffusion processes, based on the Wasserstein gradient flow formulation of the diffusion. In this formulation, the time-dependent density of the diffusion is derived as the limit of implicit Euler steps that follow the gradients of a particular free energy functional.…

Paper characterizes gradient descent in high-dimensional learning problems.

Paper improves REINFORCE for VI without restrictive assumptions.

Analytical method approximates ELBO gradient in clutter problem.

New method improves sample-efficiency in neural posterior estimation using simulator gradients.

Recent efforts on combining deep models with probabilistic graphical models are promising in providing flexible models that are also easy to interpret. We propose a variational message-passing algorithm for variational inference in such models. We make three contributions. First, we propose structured inference network…

Posterior inference in directed graphical models is commonly done using a probabilistic encoder (a.k.a inference model) conditioned on the input. Often this inference model is trained jointly with the probabilistic decoder (a.k.a generator model). If probabilistic encoder encounters complexities during training (e.g. s…

QBVI uses natural gradients for efficient Bayesian learning.

New method for constrained sampling using gradient flows.