# Balancing Reconstruction Quality and Regularisation in ELBO for VAEs

arXiv: https://arxiv.org/abs/1909.03765 · HTML: https://mlsanity.com/paper/1909.03765/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 9 Sep 2019.

## Built on by (most cited first, 9 of 9)

- [Topic Modeling as Multi-Objective Contrastive Optimization](https://mlsanity.com/paper/2402.07577/) (2024)
- [Generating Multivariate Load States Using a Conditional Variational Autoencoder](https://mlsanity.com/paper/2110.11435/) (2021)
- [Re-parameterizing VAEs for stability](https://mlsanity.com/paper/2106.13739/) (2021)
- [FedDM: Enhancing Communication Efficiency and Handling Data Heterogeneity in Federated Diffusion Models](https://mlsanity.com/paper/2407.14730/) (2024)
- [A Comprehensive Survey on Network Traffic Synthesis: From Statistical Models to Deep Learning](https://mlsanity.com/paper/2507.01976/) (2025)
- [Deep Generative and Discriminative Digital Twin endowed with Variational Autoencoder for Unsupervised Predictive Thermal Condition Monitoring of Physical Robots in Industry 6.0 and Society 6.0](https://mlsanity.com/paper/2509.12740/) (2025)
- [VSCOUT: A Hybrid Variational Autoencoder Approach to Outlier Detection in High-Dimensional Retrospective Monitoring](https://mlsanity.com/paper/2601.20830/) (2026)
- [Multipath Adaptive Gated Bottleneck Latent ODE with Raman Data Fusion for Cell Culture Process Forecasting](https://mlsanity.com/paper/2606.26520/) (2026)
- [eXact-Prior Variational Autoencoder (X-VAE): Learning Data-Adaptive Gaussian Mixture Priors for Latent Distributions](https://mlsanity.com/paper/2607.01275/) (2026)

## Abstract (arXiv)

A trade-off exists between reconstruction quality and the prior regularisation in the Evidence Lower Bound (ELBO) loss that Variational Autoencoder (VAE) models use for learning. There are few satisfactory approaches to deal with a balance between the prior and reconstruction objective, with most methods dealing with this problem through heuristics. In this paper, we show that the noise variance (often set as a fixed value) in the Gaussian likelihood p(x|z) for real-valued data can naturally act to provide such a balance. By learning this noise variance so as to maximise the ELBO loss, we automatically obtain an optimal trade-off between the reconstruction error and the prior constraint on the posteriors. This variance can be interpreted intuitively as the necessary noise level for the current model to be the best explanation of the observed dataset. Further, by allowing the variance inference to be more flexible it can conveniently be used as an uncertainty estimator for reconstructed or generated samples. We demonstrate that optimising the noise variance is a crucial component of VAE learning, and showcase the performance on MNIST, Fashion MNIST and CelebA datasets. We find our approach can significantly improve the quality of generated samples whilst maintaining a smooth latent-space manifold to represent the data. The method also offers an indication of uncertainty in the final generative model.