# Generalization Bounds For Unsupervised and Semi-Supervised Learning With Autoencoders

arXiv: https://arxiv.org/abs/1902.01449 · HTML: https://mlsanity.com/paper/1902.01449/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 4 Feb 2019.

## Built on by (most cited first, 11 of 11)

- [CROMA: Remote Sensing Representations with Contrastive Radar-Optical Masked Autoencoders](https://mlsanity.com/paper/2311.00566/) (2023)
- [Perfectly Balanced: Improving Transfer and Robustness of Supervised Contrastive Learning](https://mlsanity.com/paper/2204.07596/) (2022)
- [On PAC-Bayesian reconstruction guarantees for VAEs](https://mlsanity.com/paper/2202.11455/) (2022)
- [Contrastive Tuning: A Little Help to Make Masked Autoencoders Forget](https://mlsanity.com/paper/2304.10520/) (2023)
- [Unsupervised learning of anomalous diffusion data](https://mlsanity.com/paper/2108.03411/) (2021)
- [Multiview Representation Learning via Distributed Joint Latent Space Structuring](https://mlsanity.com/paper/2504.18455/) (2025)
- [Conv4Rec: A 1-by-1 Convolutional AutoEncoder for User Profiling through Joint Analysis of Implicit and Explicit Feedbacks](https://mlsanity.com/paper/2509.07499/) (2025)
- [Unsupervised Optimisation of GNNs for Node Clustering](https://mlsanity.com/paper/2402.07845/) (2024)
- [uaMix-MAE: Efficient Tuning of Pretrained Audio Transformers with Unsupervised Audio Mixtures](https://mlsanity.com/paper/2403.09579/) (2024)
- [Information-theoretic Generalization Analysis for VQ-VAEs: A Role of Latent Variables](https://mlsanity.com/paper/2505.19470/) (2025)
- [How Much Data Is Enough? Uniform Convergence Bounds for Generative & Vision-Language Models under Low-Dimensional Structure](https://mlsanity.com/paper/2512.23109/) (2025)

## Abstract (arXiv)

Autoencoders are widely used for unsupervised learning and as a regularization scheme in semi-supervised learning. However, theoretical understanding of their generalization properties and of the manner in which they can assist supervised learning has been lacking. We utilize recent advances in the theory of deep learning generalization, together with a novel reconstruction loss, to provide generalization bounds for autoencoders. To the best of our knowledge, this is the first such bound. We further show that, under appropriate assumptions, an autoencoder with good generalization properties can improve any semi-supervised learning scheme. We support our theoretical results with empirical demonstrations.