LieStoNet: Learning Lie Symmetries from Spatiotemporal Data for Stochastic Dynamical Systems
arXiv:2608.01582
2026
Dynamics
2 ideas extracted · analyzed Aug 31, 2026
What the math gives to ML
The paper offers a constructive mechanism for discovering continuous symmetries of stochastic dynamics: learn vector-field generators from trajectory increments while enforcing SDE determining equations, Lie-bracket closure, Lie-algebra axioms, and basis independence. The transferable asset is a differentiable procedure that turns unknown generators into measurable equivariance constraints. In neural networks, this can become a latent-space symmetry regularizer or an adaptive equivariant architecture, with quantitative diagnostics based on bracket residuals, Jacobi residuals, and the singular-value spectrum of the learned generator matrix.
Ideas from this paper
✗ Failed on benchmark
2026
Use discovered infinitesimal generators to create small continuous transformations of hidden states and force a neural predictor to commute with those transformations. This converts symmetry discovery into self-supervised augmentation without prespecifying a group, canonical coordinates, or hand-designed equivariant layers.
Useful8/10
Difficulty5/10
Novelty6/10
△ Mechanism confirmed, baseline not beaten
2026
Attach several neural vector fields to a latent representation and train them to form a closed Lie algebra rather than learning unrelated augmentation directions. The resulting generators provide data-driven continuous transformations that can be used as equivariance constraints, while bracket closure and basis-rank penalties prevent degenerate or redundant generators.
Useful8/10
Difficulty6/10
Novelty7/10