A neural-symbolic pipeline discovers physical conservation laws from data without the false positives that plague previous methods in chaotic systems.
March 24, 2026
Original Paper
From Data to Laws: Neural Discovery of Conservation Laws Without False Positives
arXiv · 2603.20474
The Takeaway
By decoupling dynamics learning from invariant discovery and using a strict constancy gate, this method correctly identifies when no law exists in chaotic systems. This solves a major reliability issue in data-driven scientific discovery (AI for Science).
From the abstract
Conservation laws are fundamental to understanding dynamical systems, but discovering them from data remains challenging due to parameter variation, non-polynomial invariants, local minima, and false positives on chaotic systems. We introduce NGCG, a neural-symbolic pipeline that decouples dynamics learning from invariant discovery and systematically addresses these challenges. A multi-restart variance minimiser learns a near-constant latent representation; system-specific symbolic extraction (p