LICENSES AND FILE SCOPE — VERSION 1.0
Copyright (c) 2026 Michael Schroeder

The author selected these licenses on 9 November 2026.
This is a component-specific license schedule, not a choice of two licenses
for the same work.

PAPERS: CC BY 4.0

The main paper, technical supplement, their TeX sources, authored figures,
and accompanying publication documentation are licensed under Creative Commons
Attribution 4.0 International (SPDX: CC-BY-4.0).
Full text: licenses/CC-BY-4.0.txt
Canonical license: https://creativecommons.org/licenses/by/4.0/
This applies to the PDFs and manuscript content within paper-source.zip.
Attribution: Michael Schroeder, Neural Computation: Optimal Pairwise-Max ReLU
Representations, Version 1.0, DOI 10.5281/zenodo.24051030.

LEAN PROOFS AND CODE: APACHE 2.0

The author's Lean proof sources, source templates, embedded finite proof data,
source decoders, Python checkers and other code are licensed under the Apache
License, Version 2.0 (SPDX: Apache-2.0). This includes the author's software and
supporting verification records within the complete Lean compendium, the code/
directory and computational inputs in paper-source.zip, and the outer verify.py.
Full text: licenses/Apache-2.0.txt
Canonical license: https://www.apache.org/licenses/LICENSE-2.0
The license applies to the unmodified archived proof sources as distributed
inside this release; no proof source needs to be edited to add a header.

THIRD-PARTY MATERIAL

Existing third-party licenses and attribution notices are preserved and are
not replaced by this schedule. External Lean and Mathlib dependencies are
not included in the compact deposit. Their existing terms apply when obtained
separately. Source comments, notices and recorded upstream provenance remain
unchanged. See THIRD-PARTY-NOTICES.txt in the outer release.
