Formal verification & math-leaning problems

This site is primarily a mathematical relativity problem database. Formal proof assistants (Lean, Coq, Isabelle, …) are most effective when statements have clear boundaries: finite-dimensional data, commuting operators, algebraic tensor identities, or separated ODEs with explicit parameters.

Orthogonal tags mark fv_suitability (high / medium / low) with a short fv_reason on every problem. These are heuristics, not judgments about intrinsic importance.

Orthogonal tags (quick reference)

Theorem status
Open / partial / conditional / solved / needs review — solution progress, orthogonal to problem type.
Problem type
Classical frontier vs quantitative sharpening vs formalization target, etc.
Research state
Literature-solved vs open-in-literature vs high-value unformalized / synthesized targets.
Family
Exact Kerr vs near-Kerr vacuum vs Kerr–Newman, Λ ≠ 0, NHEK, or broader rotating families.
Asymptotics
Asymptotically flat, de Sitter, AdS, or mixed settings (not interchangeable).
Coupling
Vacuum Einstein vs matter-coupled (e.g. Einstein–Maxwell).
Equation level
Scalar/Maxwell/linearized gravity/full Einstein, reductions, spectral operators, inverse problems.
Regime
Stationary vs dynamical; linear vs nonlinear; exterior/interior; extremal/near-extremal.
Relevance
Pure math vs mixed vs physics-facing motivation.
FV suitability
Heuristic formal-verification friendliness (high / medium / low) with a short reason on each problem.

Why some Kerr problems formalize more easily

What to defer. Full K-001-style nonlinear stability of subextremal Kerr is a global Einstein PDE program. It remains scientifically central, but it is a poor first formalization target: too many moving parts (gauge, modulation, trapping, radiation) for an initial library unless aggressively scoped to lemmas.

Best first targets

High FV suitability; emphasis on algebra, ODEs, and stationary reductions. Includes the curated track K-501–K-510.

Intermediate targets

Medium suitability: still structured, but microlocal remainders, coupled fields, or parameter uniformity add cost.

Long-term moonshots (illustrative)

Low FV suitability on this rubric — global dynamics, rough data, or open-ended modeling. Not “unimportant”; just poor first formalization objects.

Browse all problems with FV filters · Metadata definitions