# Instance-Adaptive Parametrization for Amortized Variational Inference

arXiv: https://arxiv.org/abs/2604.06796 · HTML: https://mlsanity.com/paper/2604.06796/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 19 of its 43 references that are in the map.
- First posted on arXiv 8 Apr 2026.

## Builds on (most cited first, 19 of 19)

- [Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms](https://mlsanity.com/paper/1708.07747/) (2017)
- [Variational Inference: A Review for Statisticians](https://mlsanity.com/paper/1601.00670/) (2016)
- [HyperNetworks](https://mlsanity.com/paper/1609.09106/) (2016)
- [A Comprehensive Survey of Continual Learning: Theory, Method and Application](https://mlsanity.com/paper/2302.00487/) (2023)
- [Continual learning with hypernetworks](https://mlsanity.com/paper/1906.00695/) (2019)
- [Weakly-Supervised Disentanglement Without Compromises](https://mlsanity.com/paper/2002.02886/) (2020)
- [Inference Suboptimality in Variational Autoencoders](https://mlsanity.com/paper/1801.03558/) (2018)
- [Iterative Amortized Inference](https://mlsanity.com/paper/1807.09356/) (2018)
- [Semi-Amortized Variational Autoencoders](https://mlsanity.com/paper/1802.02550/) (2018)
- [Bayesian Hypernetworks](https://mlsanity.com/paper/1710.04759/) (2017)
- [Amortized Inference Regularization](https://mlsanity.com/paper/1805.08913/) (2018)
- [Amortized Variational Inference: A Systematic Review](https://mlsanity.com/paper/2209.10888/) (2022)
- [Don't Push the Button! Exploring Data Leakage Risks in Machine Learning and Transfer Learning](https://mlsanity.com/paper/2401.13796/) (2024)
- [Strategies to exploit XAI to improve classification systems](https://mlsanity.com/paper/2306.05801/) (2023)
- [Reducing the Amortization Gap in Variational Autoencoders: A Bayesian Random Function Approach](https://mlsanity.com/paper/2102.03151/) (2021)
- [Toward the application of XAI methods in EEG-based systems](https://mlsanity.com/paper/2210.06554/) (2022)
- [Variational Hyper-Encoding Networks](https://mlsanity.com/paper/2005.08482/) (2020)
- [Semi-supervised detection of structural damage using Variational Autoencoder and a One-Class Support Vector Machine](https://mlsanity.com/paper/2210.05674/) (2022)
- [SincVAE: A new semi-supervised approach to improve anomaly detection on EEG data using SincNet and variational autoencoder](https://mlsanity.com/paper/2406.17537/) (2024)

## Abstract (arXiv)

Variational autoencoders (VAEs) rely on amortized variational inference to enable efficient posterior approximation, but this efficiency comes at the cost of a shared parametrization, giving rise to the amortization gap. We propose the instance-adaptive variational autoencoder (IA-VAE), an amortized inference framework in which a hypernetwork generates input-dependent modulations of a shared encoder. This enables input-specific adaptation of the inference model while preserving the efficiency of a single forward pass. From a theoretical perspective, we show that the variational family induced by IA-VAE contains that of standard amortized inference, implying that IA-VAE cannot yield a worse optimal ELBO. By leveraging instance-specific parameter modulations, the proposed approach can achieve performance comparable to standard encoders with substantially fewer parameters, indicating a more efficient use of model capacity. Experiments on synthetic data, where the true posterior is known, show that IA-VAE yields more accurate posterior approximations and reduces the amortization gap. Similarly, on standard image benchmarks, IA-VAE consistently improves held-out ELBO over baseline VAEs, with statistically significant gains across multiple runs. These results suggest that increasing the flexibility of the inference parametrization through instance-adaptive modulation is an effective strategy for mitigating amortization-induced suboptimality in deep generative models.