AI-developed proof of the Herzog–Schönheim conjecture for solvable groups

In an exact partition of any finite or infinite solvable group into at least two left cosets, two of the corresponding subgroups have the same index. The cardinality form of Erdős Problem 274 follows.

Sources for this record

What was already known

Several solvable subcases were known. A previous general solvable proof was withdrawn after a gap.

AI involvement AI-centered

Murali Menon chose and directed the problem. Claude agents developed and repaired most of the proof; Codex supplied a central divisibility route.

Codex supplied a central divisibility route and Claude agents built the proof, while Menon directed the problem and accepted the result; exact first authorship of every step is not logged.

Importance Notable

Extends the conjecture to all solvable groups, including infinite groups.

Limitations

The general conjecture for arbitrary groups remains open. Priority has not had specialist confirmation, and no independent human mathematical review is documented.

Technical verification Formal (Lean 4 kernel with Mathlib and Comparator)

The cited source reports sorry-free Lean proofs of the index and cardinality forms with standard axioms and Comparator statements matching the upstream exact-covering formulation plus solvability. The paper reviews were AI-only, and the machine run was not reproduced here.

Formal checking

The pinned source reports sorry-free Lean proofs of the index and cardinality forms with standard axioms and Comparator statements matching the upstream exact-covering formulation plus solvability. The paper reviews were AI-only, and the machine run was not reproduced here.

Checker
Lean kernel and Comparator in Menon's pinned repository
Scope
Erdos274.Palomar.herzog_schonheim_solvable and erdos_274_solvable for exact coverings of solvable groups
Basis
Reported by the linked checker
Evidence source

Checking a formal statement does not establish that it matches the original problem.

Human understanding Few

Default for missing human explanation: Menon's direction and acceptance do not document a person's account of the central finite-solvable argument, and the paper reviewers named here are AI systems.

Correction and publication history