注意力作为格点规范理论:关系只通过 Wilson 圈相互作用
Attention as Lattice Gauge Theory: Relations Interact Only Through Wilson Loops
AI 导读
一篇 arXiv 论文证明,若允许每个实体的坐标系独立旋转,边上注意力唯一被对称性允许的形式就是通过 Wilson 圈(holonomy $H=T_{ij}T_{jk}T_{ki}$)打分与更新。
正文 · 原文
Abstract:Relations compose through the entity they share - Alice is Bob's mother and Bob is Carol's father, so Alice is Carol's grandmother - and the order matters, since a father's mother is not a mother's father; verifying a composed relation means closing the loop. We show that if each entity's coordinate frame may be rotated independently (per-entity frame freedom), three things follow. In the pure-gauge sector, relations that share no cycle cannot interact at all. The only gauge-invariant score of a triangle is a class function of its holonomy $H = T_{ij} T_{jk} T_{ki}$ - a Wilson loop. And the only equivariant update of an edge is $U = \Psi(H)\, T_{ij}$ with $\Psi$ conjugation-equivariant. Edge attention by Wilson loops is therefore not one design among many but the unique form the symmetry permits. Three communities hold pieces of this form: the commuting-operator embeddings of knowledge graphs are its abelian (commuting) sector; AlphaFold's triangle update and the Edge Transformer have its shape - an edge updated through the third vertex - without its constraints; and lattice QCD builds it equivariantly on lattices. We instantiate the equivariant sub-family on general relational graphs as holonomy attention. With no training and composition alone, it infers unobserved $S_4$ relations on sparse graphs at 0.88 with a fifth of the edges corrupted, at or slightly below specialist synchronization algorithms, where every commutative architecture is capped at 0.083; its holonomy score flags corrupted triangles at AUC 0.99 where commuting scores are capped at 0.76 by a second selection rule, and it separates clusters without labels.
| Comments: | 7 pages, 1 figure. Accepted at the 9th Workshop on Machine Learning and the Physical Sciences (ML4PS 2026), co-located with NeurIPS 2026 |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computational Physics (physics.comp-ph) |
| Cite as: | arXiv:2610.37656 [cs.LG] |
| (or arXiv:2610.37656v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.37656 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Aakash Sarkar [view email]
[v1]
Tue, 27 Oct 2026 17:58:48 UTC (35 KB)
来源:arXiv cs.AI · arxiv.org