Cluster Algebras for Bosonic Plethysm

What the math gives to ML

The paper gives an explicit sparse, skew-symmetrizable exchange system produced by folding a larger quiver under a transpose involution. Its transferable value is the combination of sparse signed interactions, orbit-based parameter tying, and deterministic mutation schedules, rather than the plethysm application itself. These ingredients can define a positive feature-mixing block whose multiplicative-additive updates are evaluated in log space for numerical stability. The most direct test is replacing a dense MLP mixer with a fixed folded mutation circuit at matched parameter count and FLOPs.