Extend uncertainty quantification tools across self-supervised imaging methods

Determine the extent to which test-time uncertainty quantification techniques (for example, SURE-based error estimation, higher-order SURE, Tweedie-based posterior moment estimation, and equivariant bootstrapping for nullspace error) can be applied to other self-supervised learning approaches for inverse problems presented in the monograph, and characterize their validity and limitations.

Background

The manuscript discusses several uncertainty quantification approaches that can be used with self-supervised methods, including SURE/SURE-for-SURE, Tweedie's formula, and equivariant bootstrapping to estimate reconstruction error, even without ground truth.

These ideas have been demonstrated for selected scenarios, but their broader applicability to other self-supervised techniques surveyed in the manuscript has not yet been established.

References

The extent to which these ideas could be applied to other self-supervised learning solutions in this review is an interesting open problem.

— Self-Supervised Learning from Noisy and Incomplete Data  (2601.03244 - Tachella et al., 6 Jan 2026) in Chapter 5 (Extensions and open problems), Section "Uncertainty quantification and generative modelling"