Quantifying Epistemic Uncertainty in Diffusion Models

Aditi Gupta, Raphael A Meyer, Yotam Yaniv, Elynn Chen, N. Benjamin Erichson
Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, PMLR 300:3655-3663, 2026.

Abstract

To ensure high quality outputs, it is important to quantify the epistemic uncertainty of diffusion models. Existing methods are often unreliable because they mix epistemic and aleatoric uncertainty. We introduce a method based on Fisher information that explicitly isolates epistemic variance, producing more reliable plausibility scores for generated data. To make this approach scalable, we propose FLARE (Fisher-Laplace Randomized Estimator), which approximates the Fisher information using a uniformly random subset of model parameters. Empirically, FLARE improves uncertainty estimation in synthetic time-series generation tasks, achieving more accurate and reliable filtering than other methods. Theoretically, we bound the convergence rate of our randomized approximation and provide analytic and empirical evidence that last-layer Laplace approximations are insufficient for this task.

Cite this Paper

BibTeX
@InProceedings{pmlr-v300-gupta26b, title = { Quantifying Epistemic Uncertainty in Diffusion Models }, author = {Gupta, Aditi and Meyer, Raphael A and Yaniv, Yotam and Chen, Elynn and Erichson, N. Benjamin}, booktitle = {Proceedings of The 29th International Conference on Artificial Intelligence and Statistics}, pages = {3655--3663}, year = {2026}, editor = {Khan, Emtiyaz and Li, Yingzhen and Solin, Arno and Ramdas, Aaditya}, volume = {300}, series = {Proceedings of Machine Learning Research}, month = {02--05 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v300/main/assets/gupta26b/gupta26b.pdf}, url = {https://proceedings.mlr.press/v300/gupta26b.html}, abstract = { To ensure high quality outputs, it is important to quantify the epistemic uncertainty of diffusion models. Existing methods are often unreliable because they mix epistemic and aleatoric uncertainty. We introduce a method based on Fisher information that explicitly isolates epistemic variance, producing more reliable plausibility scores for generated data. To make this approach scalable, we propose FLARE (Fisher-Laplace Randomized Estimator), which approximates the Fisher information using a uniformly random subset of model parameters. Empirically, FLARE improves uncertainty estimation in synthetic time-series generation tasks, achieving more accurate and reliable filtering than other methods. Theoretically, we bound the convergence rate of our randomized approximation and provide analytic and empirical evidence that last-layer Laplace approximations are insufficient for this task. } }
Endnote
%0 Conference Paper %T Quantifying Epistemic Uncertainty in Diffusion Models %A Aditi Gupta %A Raphael A Meyer %A Yotam Yaniv %A Elynn Chen %A N. Benjamin Erichson %B Proceedings of The 29th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2026 %E Emtiyaz Khan %E Yingzhen Li %E Arno Solin %E Aaditya Ramdas %F pmlr-v300-gupta26b %I PMLR %P 3655--3663 %U https://proceedings.mlr.press/v300/gupta26b.html %V 300 %X To ensure high quality outputs, it is important to quantify the epistemic uncertainty of diffusion models. Existing methods are often unreliable because they mix epistemic and aleatoric uncertainty. We introduce a method based on Fisher information that explicitly isolates epistemic variance, producing more reliable plausibility scores for generated data. To make this approach scalable, we propose FLARE (Fisher-Laplace Randomized Estimator), which approximates the Fisher information using a uniformly random subset of model parameters. Empirically, FLARE improves uncertainty estimation in synthetic time-series generation tasks, achieving more accurate and reliable filtering than other methods. Theoretically, we bound the convergence rate of our randomized approximation and provide analytic and empirical evidence that last-layer Laplace approximations are insufficient for this task.
APA
Gupta, A., Meyer, R.A., Yaniv, Y., Chen, E. & Erichson, N.B.. (2026). Quantifying Epistemic Uncertainty in Diffusion Models . Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 300:3655-3663 Available from https://proceedings.mlr.press/v300/gupta26b.html.

Related Material