Academic paper: outline, venue, benchmark methodology

README.md's Citation section references a paper as "once published" but no
outline or venue has been decided yet. This tracks getting that started:
outline, target venue, and confirming the benchmark methodology
(cross-platform validation results already collected under
validation/results/) is presentable as-is or needs additional controlled
experiments.

Promoted from backlog after outside review (Sepand): the algorithms and
validation methodology are publication-grade, provided two things land
first, both tracked separately:

• Citation audit of the crypto modules (see the new "Citation audit" issue)
• An AI-assisted-development disclosure section suitable for a journal
  submission, distinct from CONTRIBUTING.md's dev-facing note (see the new
  "AI-assisted development disclosure section" issue)

#87 (NTT/PQClean cross-validation) is the template for the validation
rigor this paper will describe.

Start with the Citation section in README.md and the existing results under validation/results/. Use #87's NTT/PQClean cross-validation as the template for assessing the benchmark methodology, then define the paper outline and target venue. Done means recording whether the current results are presentable or identifying the controlled experiments still needed, alongside the separately tracked prerequisites.