cs.LGNov 17, 2026

Path-Aligned Constraint Equivariance

Authors: Yuhao Wang, Yuanxi Peng

Organizations: National University of Defense Technology, Changsha, China

Abstract

Learned super-resolution of physical fields must satisfy structural constraints---such as divergence-free flow (∇⋅u=0\nabla\cdot u=0), curl-free fields (∇×u=0\nabla\times u=0), and pointwise sum-to-one normalization (1⊤p=11^\top p=1)---that standard training objectives do not enforce. Soft penalties trade fidelity against satisfaction via a scalar hyperparameter λλ but cannot drive the residual to zero; hard-constrained regression eliminates the residual by construction yet converges to the conditional mean, losing small-scale spectral structure. We observe that in Conditional Flow Matching with linear paths, a linear constraint Ax=cAx=c is preserved along the entire transport trajectory whenever the velocity field is compatible (Avθ=0Av_θ=0) and the initial noise is consistent (Ax0=cAx_0=c); along the trajectory, the latter two conditions are also necessary. We term this the path-equivariance principle and instantiate it in PACE (Path-Aligned Constraint Equivariance), a framework that enforces these conditions through a null-space parameterization of the velocity field and a constraint-respecting noise sampler, with no modification to the loss function or the ODE integrator. Across three constraint families, PACE reduces the mean constraint residual by one to five orders of magnitude relative to unconstrained baselines, matching the residual level of the data-generation pipeline itself. The proposed framework requires only a closed-form parameterization onto the null space of AA; within this scope, it offers exact constraint enforcement while recovering the spectral content and predictive diversity that conditional-mean regression systematically loses, all under a single unified recipe.

Figures & tables

Explore similar work

CardsList
  1. Difficulty-Calibrated Interpolation Paths for Conditional Flow Matching

    Aug 21, 2026Airin Akter Tania, Md Raihan KhanFlow MatchingConditional Flow Matching

  2. MintFlow: Minimal Trajectory Intervention for Constrained Flow Matching

    Sep 30, 2026Yesom Park, Kelvin Kan, Qifan Chen +3constrained flow matchingConstrained Generative Modeling

  3. Co-Evolving Paths and Flows via Path-Flow Alignment

    Oct 6, 2026Zeyu Michael Li, William Xingxu Chen, Xiang ChengFlow Matching