Observation-Aligned Mask Priors for Learning Physical Dynamics from Authentic Occlusions
For researchers working on physical dynamics from sparse sensor data (e.g., oceanography), this provides a principled way to handle authentic occlusions, though the improvement is incremental over existing diffusion methods.
The paper addresses learning physical dynamics from incomplete observations with authentic occlusions, proposing Observation-Aligned Mask Priors that learn real occlusion distributions via a Bayesian Flow Network. This method consistently improves MSE and PSNR over diffusion baselines on three real-world oceanographic datasets up to 256×256 resolution.
Learning physical dynamics directly from incomplete observations is challenging because authentic occlusions are structured, sample-dependent, and often missing not at random, whereas existing methods typically rely on heuristic masking rules or predefined mask distributions. We propose Observation-Aligned Mask Priors, a framework that learns the distribution of authentic observation masks and uses it to construct context-query partitions for training from incomplete data. Specifically, we pretrain a Bayesian Flow Network (BFN) on binary observation masks to capture real occlusion topologies, then guide BFN sampling with a globally normalized cross-entropy objective to generate sample-specific masks aligned with each sparse observation. The intersection between the guided mask and the observed mask defines the context, and the remaining observed entries become query targets for a diffusion-based reconstruction model. We show that this intersection-based partitioning gives every valid observed dimension a strictly positive probability of being queried, preventing zero-query dead zones and local generative collapse. Experiments on three real-world oceanographic datasets with authentic satellite occlusions, across resolutions up to 256$\times$256, show consistent improvements over strong diffusion baselines in MSE and PSNR. These results demonstrate that learning mask priors from authentic occlusions is an effective alternative to heuristic masking for learning from incomplete physical observations without access to fully observed fields.