arXiv Science⌕ Search

arXiv · 2611.16468

Physics-constrained differentiable causal discovery of constitutive mechanisms through boundary value problems

Abstract

Data-driven constitutive model discovery is recast as causal discovery of mechanism parent sets using time-unrolled structural causal models. Hard-coding balance laws and constitutive skeletons, candidate graphs score environment--mechanism and state--state edges via weak-form equilibrium gaps evaluated from full-field displacements, avoiding the need to measure stresses or internal variables. An adaptive lasso and a NOTEARS-type constraint ensure graph acyclicity and an explicit evaluation order, treating climate-chamber protocols as interventions. Automatic differentiation yields the score's Gauss--Newton matrix as a local identifiability certificate, driving greedy D-optimal test design. In a synthetic study on a hygro-thermo-viscoelastic damaging plate (8 true edges among 13 candidates), interventional data recover the exact true graph in 7 of 10 runs (0.01px displacement noise) with zero attribution error. Conversely, observational data with a temperature--moisture correlation of $-0.998$ fail entirely because confounded scores are 250--380 times flatter. Observationally discovered models reproduce training climates (3.6% error) but mispredict hot and humid climates by up to 41.5% (16.9% average), whereas interventionally discovered models stay within 3.2%. A correlation sweep shows the certificate successfully predicts attribution errors, though weak-form score noise bias remains a primary limitation. Computed without knowledge of the truth, the D-optimal design selects test sets ranked second or third out of 126 candidates.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Michael A. Kraus. 2026-11-12. Physics-constrained differentiable causal discovery of constitutive mechanisms through boundary value problems. https://arxiv.science/abs/2611.16468

Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite arXiv Science by Indiciae: BibTeX · RIS.

KEEP EXPLORING

Related papers

Absence of ground states for anions

We show that the $N$-electron Hamiltonian $H(N, Z)$ with the total nuclear charge $Z$ has no normalizable ground state if the ground state energy $E(N, Z)$ satisfies $E(N, Z)= E(N-1, Z)$ for $Z=N-1$. For anions $\mathrm{He}^-, \mathrm{Be}^-, \mathrm{N}^-, \mathrm{Ne}^-$, etc., many numerical results give strong evidence of the condition $E(N, Z)= E(N-1, Z)$.

math-ph↗

Cluster Expansion and Decay of Correlations for Multidimensional Long-Range Ising Models

We develop the cluster expansion for the multidimensional multiscaled contours defined by three of us. These contours are suitable for long-range Ising models with interaction $J_{xy}=J(|x-y|)= J/|x-y|^α$, $J>0$, and $α>d$. As an application of the convergence of the cluster expansion at low temperatures, we study the decay of the truncated two-point correlation functions, showing that the decay is algebraic with coefficient $α$.

math-ph↗

A Shape-Variational Characterisation of Multi-region Relaxed Magnetohydrodynamic Equilibria

Let $Λ\subset\mathbb{R}^3$ be a region admitting a partition into $n$ compact, connected subregions $Λ_1,\dots,Λ_n$, each with smooth boundary such that the interfaces between regions are closed and embedded in the interior $\text{int}(Λ)$. Consider a vector field $B$ on $Λ$ where $B|_{Λ_i}$ is smooth, divergence free, and tangent to $\partial Λ_i$ for all $i$. For each such partition we show that the multi-region relaxed magnetohydrodynamics (MRxMHD) equilibrium equations are necessary and sufficient conditions for the magnetic field, together with the placement of the interfaces, to be a stationary point of the magnetic energy under appropriate constraints. We prescribe the pressure, loop-corrected helicity, and magnetic flux of $B$ through all smooth surfaces in $Λ_i$ whose boundary lies on $\partial Λ_i$. We isolate a gauge condition that is imposed implicitly in the existing toroidal treatments and show that the variation of the loop-corrected helicity cancels the corresponding boundary contribution without restricting the potentials. On toroidal domains this helicity agrees with the familiar loop-corrected expressions. For a single fixed region we give a condition on the Beltrami parameter of a region that is sufficient for a critical point of the magnetic energy to be the strict global minimiser of the energy among all admissible fields of that region. This is a condition for Beltrami fields on fixed domains that is applicable to individual regions $Λ_i$ given a fixed partition and metric.

math-ph↗