UQ-SHRED: uncertainty quantification of shallow recurrent decoder networks for sparse sensing via engression

View PDF HTML (experimental)

Abstract:Reconstructing high-dimensional spatiotemporal fields from sparse sensor measurements is critical in a wide range of scientific applications. The SHallow REcurrent Decoder (SHRED) architecture is a recent state-of-the-art architecture that reconstructs high-quality spatial domain from hyper-sparse sensor measurement streams. An important limitation of SHRED is that in complex, data-scarce, high-frequency, or stochastic systems, portions of the spatiotemporal field must be modeled with valid uncertainty estimation. We introduce UQ-SHRED, a distributional learning framework for sparse sensing problems that provides uncertainty quantification through a neural network-based distributional regression called engression. UQ-SHRED models the uncertainty by learning the predictive distribution of the spatial state conditioned on the sensor history. By injecting stochastic noise into sensor inputs and training with an energy score loss, UQ-SHRED produces predictive distributions with minimal computational overhead, requiring only noise injection at the input and resampling through a single architecture without retraining or additional network structures. On complicated synthetic and real-life datasets including turbulent flow, atmospheric dynamics, neuroscience and astrophysics, UQ-SHRED provides a distributional approximation with well-calibrated confidence intervals. We further conduct ablation studies to understand how each model setting affects the quality of the UQ-SHRED performance, and its validity on uncertainty quantification over a set of different experimental setups.

Subjects:

Machine Learning (cs.LG); Computational Engineering, Finance, and Science (cs.CE)

Cite as: arXiv:2604.01305 [cs.LG]

(or arXiv:2604.01305v1 [cs.LG] for this version)

https://doi.org/10.48550/arXiv.2604.01305

arXiv-issued DOI via DataCite

Submission history

From: Yuxuan Bao [view email] [v1] Wed, 1 Apr 2026 18:18:09 UTC (21,476 KB)