Semigroup Consistency as a Diagnostic for Learned Physics Simulators

Lennon J. Shikhman

cs.LGMay 25, 2026

Semigroup Consistency as a Diagnostic for Learned Physics Simulators

Authors: Lennon J. Shikhman

Abstract

Learned physics simulators are often evaluated by one-step or short-horizon prediction error, but these metrics can miss failures in temporal composition and long-horizon rollout. For autonomous, state-complete systems, exact solution maps satisfy a semigroup law: direct evolution over $s+t$ should agree with evolution over $s$ followed by $t$ . We propose normalized semigroup error as a post hoc, model-agnostic diagnostic comparing these direct and composed learned predictions. On one-dimensional heat and Burgers dynamics with time-conditioned ConvNet and FNO baselines, semigroup error is positively associated with rollout degradation, with trajectory-level Spearman correlation $ρ= 0.635$ and $95%$ CI $[0.621, 0.649]$ . Semigroup regularization has mixed effects, supporting semigroup consistency primarily as an evaluation diagnostic rather than a universally beneficial training objective.

Explore similar work

Semigroup-JEPA: Latent Dynamics Consistency for Zero-Shot Physics Generalization
Sep 9, 2026Andy Zeyi Liu, Haoran Sun, Lucas Baker +2Joint-Embedding Predictive Architectures Latent World Models
Prediction Is Not Physics: Learning and Evaluating Conserved Quantities in Neural Simulators
May 16, 2026Andrew Bukowski, Aditya Kothari, Simba Shi +1Neural Ordinary Differential Equations Neural Control
Conservation Buys Stability and Factoring Buys Counterfactuals in Physical World Models
Sep 17, 2026Yufeng Wang, Parivesh Priye, Lu Wei +1Counterfactual Reasoning World Models

Sep 9, 2026 · cs.LGJ/K move · Enter open · S save

Semigroup-JEPA: Latent Dynamics Consistency for Zero-Shot Physics Generalization

Andy Zeyi Liu, Haoran Sun, Lucas Baker, Randall Balestriero+1

Joint-Embedding Predictive Architecture (JEPA) world models learn a compact latent representation of the world that supports prediction and planning, but their capability to learn physics and generate physically realistic dynamics remains hitherto untested. In this work, we introduce SemiGroup-JEPA (SG-JEPA), which extends the LeWorldModel framework by supplying the parameter governing the physics to the temporal model via action-conditioning and jointly training an encoder and predictor through an autoregressive latent rollout. To evaluate the model's ability to generalize out of distribution, we design dynamical tasks under different gravitational fields that, despite obeying the same physical law, exhibit qualitatively different dynamics, ranging from floating motion in weak gravitational fields to rapid bouncing in strong ones. In contrast to DINO-WM, SG-JEPA reduces open-loop prediction error by up to 2 times on two-dimensional datasets, and increases control success rate up to 2.5 times for three-dimensional robotic datasets, for which we train independent diffusion policies. To explain this advantage, we develop a linear feature model that separates local law-conditioned error from its recursive amplification under rollout. Guided by this model, we find that back-propagating the multi-step rollout loss into the representation trains the encoder to keep the features that the predictor can carry forward, and that those are the features the dynamics depend on, so most of the gain comes from the encoder learning better features rather than from the predictor learning better dynamics. See project page at https://sg-jepa.github.io.

Joint-Embedding Predictive Architectures Latent World Models Sphere-Jepa Latent Dynamics Autoregressive Rollout Robotic Datasets

Open paper Ask about it PDF

Abstract

Explore similar work

Semigroup-JEPA: Latent Dynamics Consistency for Zero-Shot Physics Generalization

Prediction Is Not Physics: Learning and Evaluating Conserved Quantities in Neural Simulators

Conservation Buys Stability and Factoring Buys Counterfactuals in Physical World Models