Empirical Variational Autoencoder

Hugging Face Daily Papers Papers
variational-autoencoder generative-models sequence-modeling autoregressive-priors image-synthesis sound-generation

Summary

Empirical Variational Autoencoder (EVA) is a general generative framework for continuous-valued sequences that learns autoregressive latent priors empirically from training data, alleviating the prior-posterior gap of conventional VAEs and achieving competitive generation quality with autoregressive diffusion baselines at much faster inference speed.

We present Empirical Variational Autoencoder, a general generative framework for continuous-valued (i.e., non-vector-quantized) sequences. EVA is based on the evidence lower bound of the Variational Autoencoder (VAE) but learns autoregressive latent priors empirically from training data, which can be implemented only by an additional single linear layer on top of VAEs. By replacing the conventional standard-Gaussian constraint with the self-predicted priors, EVA significantly alleviates the latent distribution gap between prior and posterior which is typically observed in conventional VAEs, and leads to high-fidelity ancestral sampling for sequential data generation. Extensive experiments on image and sound synthesis demonstrate that EVA achieves competitive generation quality with autoregressive diffusion baselines despite its much faster inference time.
Original Article
View Cached Full Text

Cached at: 10/06/26, 04:57 PM

Paper page - Empirical Variational Autoencoder

Source: https://huggingface.co/papers/2610.06545

Abstract

WepresentEmpiricalVariationalAutoencoder,ageneralgenerativeframeworkforcontinuous-valued(i.e.,non-vector-quantized)sequences.EVAisbasedontheevidencelowerboundoftheVariationalAutoencoder(VAE)butlearnsautoregressivelatentpriorsempiricallyfromtrainingdata,whichcanbeimplementedonlybyanadditionalsinglelinearlayerontopofVAEs.Byreplacingtheconventionalstandard-Gaussianconstraintwiththeself-predictedpriors,EVAsignificantlyalleviatesthelatentdistributiongapbetweenpriorandposteriorwhichistypicallyobservedinconventionalVAEs,andleadstohigh-fidelityancestralsamplingforsequentialdatageneration.ExtensiveexperimentsonimageandsoundsynthesisdemonstratethatEVAachievescompetitivegenerationqualitywithautoregressivediffusionbaselinesdespiteitsmuchfasterinferencetime.

View arXiv pageView PDFProject pageGitHub0Add to collection

Get this paper in your agent:

hf papers read 2610\.06545

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.06545 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.06545 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.06545 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

ELVAE: Evidential Learning-Based Variational Autoencoder for Uncertainty-Aware Generation

arXiv cs.LG

Introduces ELVAE, a VAE with evidential learning that models latent coordinates with normal-inverse-gamma posteriors to obtain explicit uncertainty estimates. Experiments on MNIST show that within-class uncertainty ranking can stratify synthetic sample reliability and stress-testing, though results require class-wise normalization and vary across seeds.

Variational lossy autoencoder

OpenAI Blog

OpenAI researchers present a Variational Lossy Autoencoder (VLAE) that combines VAEs with neural autoregressive models (RNN, MADE, PixelRNN/CNN) to learn controllable global representations, achieving state-of-the-art results on MNIST, OMNIGLOT, and Caltech-101 Silhouettes density estimation tasks.

Continuous First, Discrete Later: VQ-VAEs Without Dimensional Collapse

arXiv cs.LG

This paper addresses the issue of dimensional collapse in VQ-VAEs, showing that representations often occupy a low-dimensional subspace. It proposes an 'AE Warm-Up' strategy that trains the model as an unquantized autoencoder first, which improves reconstruction quality and increases effective latent dimensionality.

V-RAE: Rethinking Video Latent Spaces for Generation

Hugging Face Daily Papers

V-RAE proposes a video representation autoencoder that builds semantically organized latents from frozen vision representations to enhance video generation quality, convergence speed, and predictive modeling.