Machines find the hidden language of physical measurements
Discovering Physical Representation Languages
Machine LearningArtificial Intelligence
Summary
Before machines can find physical laws, they need to understand what their measurement data actually represent. The authors develop a method that helps machines discover this hidden measurement language just by looking at controlled experiments, without prior knowledge. Their method works efficiently and can handle complicated situations, even when the data is noisy or incomplete. This advances the ability of machines to learn physical laws in the languages where these laws naturally make sense rather than in human-made formulas.
What this means in practice
- •For experimental physics teams: Identify and interpret complex physical measurement structures from controlled experiments without prior labeling or assumptions about data types.
- •For industrial sensor engineers: Develop robust sensor data interpretation systems that adjust to nonlinearities, memory effects, and noise by discovering underlying measurement representations automatically.
Authors
Linzhe Zhang, Changming Xu
Abstract
Before a machine can discover a physical law, it must discover what its measurements are: which observations live on cells, which are intensive or extensive, which sectors are dual, and which distinctions are merely gauge. We introduce physical representation-language discovery, the problem of recovering this hidden ontology directly from anonymous controlled experiments. We give an identifiability theory and constructive polynomial-time procedure that recovers a carrier and differential sequence, measurement types and orientation twist, noninvertible refinement semantics, primal-dual Maxwell diagrams, and the residual equivalences that no permitted experiment can break. The theory turns material nuisance into a commutant, uses refinement to separate quantities from coordinates, and selects physics only after its representation has been recovered. For a certified finite experiment family, we prove an end-to-end two-stage measurement bound and a matching minimax rate in dimension, accuracy, and confidence. Blind Maxwell experiments recover complete primal/relative-dual ontologies on regular and unstructured carriers under jointly corrupted observations; an independent unstructured RLC system demonstrates that the result is not specific to Maxwell. The framework scales to tens of thousands of cells per carrier, while stress audits demonstrate robustness across severe physical regimes - including non-Markovian memory, nonlinearities, nonlocality, and complex constitutive hysteresis. A public FDTD audit demonstrates the emergence of anonymous curl structure from incomplete field data, while characterizing the informational prerequisites for complete recovery. The goal is to move scientific ML from learning laws in a human-supplied language to discovering the language in which laws become expressible, establishing exact theoretical limits on observational identifiability.