OpenAI preprint, September 2026 · §2. Rectangular tests from subspace expansion · §2
§2 (setting): the reduction of a Laurent-series subspace:Integral vectors of a subspace
of
,
reduced entrywise to their constant coefficients, and the
-subspace
of
they span
The reduction modulo
of a subspace over
..
Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its
declaration both in Lean and in words.
The Setting
Let be a field, an index set and a subspace of the space of vectors indexed by whose entries are formal Laurent series in over , spanned for instance by the columns of a matrix with such entries. What is the value of at ? A spanning vector with an entry has no value there, and putting into two spanning sets of the same can give two different subspaces of .
The library’s answer is the reduction of , a subspace of , defined for vectors indexed by any set .
The Definition
Definition 1. Let be a field. The Laurent series field over is the field of formal Laurent series in one variable with coefficients in ,(1)✦A Laurent series in is integral when it lies in the power series ring :(2)✦The reduction of a Laurent series in is its constant coefficient .✦For an index set , a vector in is integral when every entry is integral✦, and its reduction is the vector in .✦For an -linear subspace of , the reduced set of is the subset of formed by the reductions of the integral vectors of :(3)✦The reduction of is the -linear subspace of spanned by the reduced set: (4)✦
In the paper, the result reads: with , and the paper’s index set for the entries of vectors, the reduction of a subspace is , taken over all integral vectors of .
What Is Checked
The definitions (1) to (4) are stated over an arbitrary field , an arbitrary index set and an -linear subspace of , with integrality and reduction modulo of a vector taken entry by entry.