Cyclic Shuffle Groups: Universal Two-Transitivity and Complete Classification
arXiv:2608.12526
2026
Architecture
1 ideas extracted · analyzed Sep 1, 2026
What the math gives to ML
The paper studies two explicit permutations on N=kn positions: a perfect shuffle and a cyclic pile rotation. Its transferable asset is the constructive group action: when n is not a power of k, these two O(N)-cost reindexings generate a 2-transitive group, so every ordered pair of positions can in principle be transported to every other ordered pair. This suggests sparse token-mixing architectures that replace dense N-by-N attention with repeated fixed permutations plus learned channel transformations. The theorem guarantees pair reachability, while word length, optimization behavior, and practical information mixing must be tested empirically.
Ideas from this paper
Unverified
2026
Construct a sparse token-mixing architecture by interleaving learned per-token transformations with fixed perfect-shuffle and cyclic-pile permutations. For n not a power of k, the generated permutation group is 2-transitive, so sufficiently rich sequences of generator words can expose every ordered token pair without constructing a dense N by N attention matrix.
Useful5/10
Difficulty4/10
Novelty6/10