Zesting and the relative complexity of Reshetikhin-Turaev invariants
arXiv:2608.02795
2026
Architecture
1 ideas extracted · analyzed Aug 31, 2026
What the math gives to ML
The paper develops zesting, a controlled modification of an A-graded ribbon fusion category using invertible objects, associator maps, braiding maps, and twist factors. The transferable asset is a finite-group-graded local interaction rule whose coherence constraints make multi-step composition independent of parenthesization. A neural analogue is a group-graded attention or message-passing layer with cocycle-constrained interaction phases or signs, replacing unconstrained pairwise gates with algebraically consistent composition. This is most promising when tokens or graph nodes have known latent group labels and the task requires multi-hop relational reasoning.
Ideas from this paper
Unverified
2026
Attach each token or graph node a discrete grade a in a finite group A, and modify attention value composition with a normalized group 2-cocycle rather than independent pairwise gates. The cocycle provides a globally consistent projective interaction rule, so composing three messages gives the same result under either parenthesization. This may improve relational reasoning while reducing the number of freely learned interaction parameters.
Useful5/10
Difficulty5/10
Novelty7/10