ScaleAware-JEPA: Latent Representation for Discovery in Multiscale Physical Fields
2026-06-29 • Machine Learning
Machine LearningComputer Vision and Pattern Recognition
AI summaryⓘ
The authors address the challenge of analyzing complex physical data that change at different scales, where usual methods rely on fixed grid positions that don't match the data's natural structure. They propose ScaleAware-JEPA, a new approach that breaks down data into parts based on scale and uses this to better predict and understand the hidden patterns without using labels. Their method works on various scientific data sets and creates detailed maps of structures that don't require pre-set categories. This helps scientists explore complex phenomena in a more flexible, scale-aware way.
self-supervised learninglatent coordinatesscale hierarchyConstrained Diffusion Decompositionphysical fieldsmultiscale structurejoint-embedding predictive architecture (JEPA)MHD turbulencemolecular gasurban nighttime-light data
Authors
Guang-Xing Li
Abstract
Continuous physical fields represent a large fraction of data under scientific investigation. Their multiscale structures are central to discovery, yet useful coordinates are not known in advance. Standard self-supervised methods define context and targets in fixed image coordinates, posing a predictive task misaligned with fields organized across a continuous scale hierarchy. We introduce ScaleAware-JEPA, a framework that constructs dense, label-free latent coordinates for continuous scalar fields. Constrained Diffusion Decomposition (CDD) separates each field into pixel-registered scale components and provides the scale coordinates that define the masking geometry. The resulting JEPA objective predicts hidden structure with a context footprint tied to the diffusion scale of each component rather than to an arbitrary patch size. Across MHD turbulence, interstellar molecular gas and urban nighttime-light structure, the learned geometry maps back to coherent morphology, forming dense structural atlases without labels or predefined segmentation rules. By tying latent prediction to the scale hierarchy of a field, ScaleAware-JEPA constructs latent coordinates through which complex physical patterns can be inspected before their relevant structures have been prescribed. Code is available at https://github.com/gxli/SA-JEPA.