This paper by Xin Huang proposes a clean translation of a conjecture from group theory into the language of finite categories and shows the two versions are actually equivalent. The original statement, called the Galois Alperin weight conjecture (GAWC), predicts a natural match between two kinds of basic building blocks (simple modules) for a finite group algebra, where the match respects the action of the Galois group of the field of p elements. Huang formulates the same prediction for a finite category and proves that nothing new is gained: the category version holds exactly when the group version does.

At a high level, the objects of interest are algebras built from a finite category C over a field k that is an algebraic closure of the finite field F_p, and the set Γ of field automorphisms (the Galois group). One builds a related category called the p-orbit category O_C, introduced earlier by Linckelmann. The conjecture says there should be a Γ-equivariant bijection between the simple modules of the category algebra kC and the simple modules of a certain ‘‘weight algebra’’ W(kO_C). Informally, weights are pieces of representation data attached to idempotent endomorphisms inside C and to simple modules of certain groups that appear inside the category.

Huang works with both ordinary and twisted category algebras (these twists are governed by 2-cocycles, a kind of bookkeeping for multiplication signs). A key part of the paper is to show that if the expected Γ-equivariant bijections exist for the group-like building blocks attached to each idempotent, then they assemble to give the bijection for the whole category algebra. This yields the main equivalence theorems: the GAWC for finite groups is equivalent to the category version, and a blockwise refinement (which organizes representations into pieces called blocks) is equivalent for finite EI-categories. EI-categories are ones where every endomorphism is actually an isomorphism; they include ordinary groups as a special case.