Skip to content

Package S10 tail assumptions for reusable callers#31

Open
freezed-corpse-143 wants to merge 1 commit into
rm-lieb-s10-trace-mgf-to-tail-assumption-compositionfrom
rm-lieb-s10b-tail-assumption-bundle-contract
Open

Package S10 tail assumptions for reusable callers#31
freezed-corpse-143 wants to merge 1 commit into
rm-lieb-s10-trace-mgf-to-tail-assumption-compositionfrom
rm-lieb-s10b-tail-assumption-bundle-contract

Conversation

@freezed-corpse-143

Copy link
Copy Markdown
Collaborator

Summary

  • Add a reusable MatrixBernsteinConditioningTraceMGFTailAssumptions bundle for the S10 conditioning/trace-MGF-to-tail route.
  • Expose matrixBernsteinQuadraticFormUpperTail_of_conditioningTraceMGFTailAssumptions as a downstream-friendly wrapper with honest assumption boundaries.
  • Update API/judge smoke checks and roadmap docs so callers can depend on the S10 tail-assumption bundle explicitly.

Motivation

This advances the aggressive Matrix Bernstein/Lieb stack by packaging the S10 tail route behind a reusable assumption contract instead of requiring each downstream caller to thread the conditional trace-MGF assumptions manually.

Scope

  • Adds an assumption-bundle API and wrapper in HighDimProb/RandomMatrix/ConcentrationStatements.lean.
  • Extends public API and judge checks in HighDimProbTest/RandomMatrixConcentrationAPI.lean and HighDimProbJudge/RandomMatrix/MatrixBernsteinUse.lean.
  • Updates RandomMatrix roadmap/status docs to record the new contract and preserved assumption boundary.
  • Does not prove new analytic primitives beyond the stated assumption contract.

Testing

  • git diff --check
  • python .github/scripts/check_text_quality.py
  • python scripts/judge_policy_check.py
  • lake build HighDimProb.RandomMatrix.ConcentrationStatements
  • lake env lean HighDimProbTest/RandomMatrixConcentrationAPI.lean
  • lake env lean HighDimProbJudge/RandomMatrix/MatrixBernsteinUse.lean
  • lake build HighDimProbJudge
  • lake test

Constraint: Keep theorem boundaries honest by bundling only explicit S9 conditioning and tail bridge premises.
Rejected: Proving new conditioning or tail-domination facts | task requires no new hard facts.
Confidence: high
Scope-risk: narrow
Directive: Do not treat this bundle as a full or unconditional Matrix Bernstein theorem.
Tested: lake build HighDimProb.RandomMatrix.ConcentrationStatements; lake env lean HighDimProbTest/RandomMatrixConcentrationAPI.lean; lake env lean HighDimProbJudge/RandomMatrix/MatrixBernsteinUse.lean; python .github/scripts/check_text_quality.py; python scripts/judge_policy_check.py; git diff --check
Not-tested: lake test
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant