# The Variational Homoencoder: Learning to learn high capacity generative models from few examples

arXiv: https://arxiv.org/abs/1807.08919 · HTML: https://mlsanity.com/paper/1807.08919/

## Key facts

- Cited by 23 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 24 Jul 2018.

## Built on by (most cited first, 23 of 23)

- [Neural ODE Processes](https://mlsanity.com/paper/2103.12413/) (2021)
- [Few-Shot Diffusion Models](https://mlsanity.com/paper/2205.15463/) (2022)
- [The Neural Process Family: Survey, Applications and Perspectives](https://mlsanity.com/paper/2209.00517/) (2022)
- [Zero-shot Imputation with Foundation Inference Models for Dynamical Systems](https://mlsanity.com/paper/2402.07594/) (2024)
- [Recursively Conditional Gaussian for Ordinal Unsupervised Domain Adaptation](https://mlsanity.com/paper/2107.13467/) (2021)
- [In-Context Learning of Stochastic Differential Equations with Foundation Inference Models](https://mlsanity.com/paper/2502.19049/) (2025)
- [SCHA-VAE: Hierarchical Context Aggregation for Few-Shot Generation](https://mlsanity.com/paper/2110.12279/) (2021)
- [Few-shot Learning for Topic Modeling](https://mlsanity.com/paper/2104.09011/) (2021)
- [Drawing out of Distribution with Neuro-Symbolic Generative Models](https://mlsanity.com/paper/2206.01829/) (2022)
- [Learning multi-modal generative models with permutation-invariant encoders and tighter variational objectives](https://mlsanity.com/paper/2309.00380/) (2023)
- [Latent Representation Matters: Human-like Sketches in One-shot Drawing Tasks](https://mlsanity.com/paper/2406.06079/) (2024)
- [Meta-Statistical Learning: Supervised Learning of Statistical Estimators](https://mlsanity.com/paper/2502.12088/) (2025)
- [Consolidated learning -- a domain-specific model-free optimization strategy with examples for XGBoost and MIMIC-IV](https://mlsanity.com/paper/2201.11815/) (2022)
- [Computing with Categories in Machine Learning](https://mlsanity.com/paper/2303.04156/) (2023)
- [Towards Generative Abstract Reasoning: Completing Raven's Progressive Matrix via Rule Abstraction and Selection](https://mlsanity.com/paper/2401.09966/) (2024)
- [Learning to Infer Generative Template Programs for Visual Concepts](https://mlsanity.com/paper/2403.15476/) (2024)
- [Meta-learning for Matrix Factorization without Shared Rows or Columns](https://mlsanity.com/paper/2106.15133/) (2021)
- [Meta-learning One-class Classifiers with Eigenvalue Solvers for Supervised Anomaly Detection](https://mlsanity.com/paper/2103.00684/) (2021)
- [OR-Net: Pointwise Relational Inference for Data Completion under Partial Observation](https://mlsanity.com/paper/2105.00397/) (2021)
- [Meta-learning Amidst Heterogeneity and Ambiguity](https://mlsanity.com/paper/2107.02228/) (2021)
- [A Probabilistic Generative Model of Free Categories](https://mlsanity.com/paper/2205.04545/) (2022)
- [Character Generation through Self-Supervised Vectorization](https://mlsanity.com/paper/2208.02012/) (2022)
- [Human-machine knowledge hybrid augmentation method for surface defect detection based few-data learning](https://mlsanity.com/paper/2304.13963/) (2023)

## Abstract (arXiv)

Hierarchical Bayesian methods can unify many related tasks (e.g. k-shot classification, conditional and unconditional generation) as inference within a single generative model. However, when this generative model is expressed as a powerful neural network such as a PixelCNN, we show that existing learning techniques typically fail to effectively use latent variables. To address this, we develop a modification of the Variational Autoencoder in which encoded observations are decoded to new elements from the same class. This technique, which we call a Variational Homoencoder (VHE), produces a hierarchical latent variable model which better utilises latent variables. We use the VHE framework to learn a hierarchical PixelCNN on the Omniglot dataset, which outperforms all existing models on test set likelihood and achieves strong performance on one-shot generation and classification tasks. We additionally validate the VHE on natural images from the YouTube Faces database. Finally, we develop extensions of the model that apply to richer dataset structures such as factorial and hierarchical categories.