# General Uncertainty Estimation with Delta Variances

arXiv: https://arxiv.org/abs/2502.14698 · HTML: https://mlsanity.com/paper/2502.14698/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 52 references that are in the map.
- First posted on arXiv 20 Feb 2025.

## Builds on (most cited first, 5 of 5)

- [Understanding Black-box Predictions via Influence Functions](https://mlsanity.com/paper/1703.04730/) (2017)
- [Optimizing Neural Networks with Kronecker-factored Approximate Curvature](https://mlsanity.com/paper/1503.05671/) (2015)
- [Out-of-Distribution Detection with Deep Nearest Neighbors](https://mlsanity.com/paper/2204.06507/) (2022)
- [A Simple Baseline for Bayesian Uncertainty in Deep Learning](https://mlsanity.com/paper/1902.02476/) (2019)
- [Learning Expressive Priors for Generalization and Uncertainty Estimation in Neural Networks](https://mlsanity.com/paper/2307.07753/) (2023)

## Built on by (most cited first, 3 of 3)

- [Uncertainty and Fairness Awareness in LLM-Based Recommendation Systems](https://mlsanity.com/paper/2602.02582/) (2026)
- [An Isotropic Approach to Efficient Uncertainty Quantification with Gradient Norms](https://mlsanity.com/paper/2603.29466/) (2026)
- [Tracing sources of epistemic uncertainty in deep learning predictions: homo- and hetero-scedastic linearized estimators](https://mlsanity.com/paper/2608.07630/) (2026)

## Abstract (arXiv)

Decision makers may suffer from uncertainty induced by limited data. This may be mitigated by accounting for epistemic uncertainty, which is however challenging to estimate efficiently for large neural networks. To this extent we investigate Delta Variances, a family of algorithms for epistemic uncertainty quantification, that is computationally efficient and convenient to implement. It can be applied to neural networks and more general functions composed of neural networks. As an example we consider a weather simulator with a neural-network-based step function inside -- here Delta Variances empirically obtain competitive results at the cost of a single gradient computation. The approach is convenient as it requires no changes to the neural network architecture or training procedure. We discuss multiple ways to derive Delta Variances theoretically noting that special cases recover popular techniques and present a unified perspective on multiple related methods. Finally we observe that this general perspective gives rise to a natural extension and empirically show its benefit.