Plectis Atlas
Search

↑ and ↓ move, Enter opens, Esc closes.

Lean statement

FPRAS for common matroid bases

Lean is a language for mathematical statements and proofs that a computer can verify. This page shows one statement file from OpenAI’s release, possible paper matches and independent reports. The code-like names below identify the definitions or theorems selected by the proof-verification settings.

Theorem selected for verification: OAI.common_bases_fpras

Lean statement file: CommonBasesFPRAS.leanShow on the mapFamily 114. OpenAI lists this as a main result in its Lean catalogue.

The Lean statement

The lines of the Lean statement that state the theorem, then the definitions in the file that it uses.

def MainTheorem : Prop :=  ∃ (k s : ℕ) (P : Program k s) (C d : ℕ), 0 < C ∧    ∀ (n r : ℕ) (M₁ M₂ : Matroid (Fin n)) (O : IndependenceOracle n)      (ε δ : ℚ),      M₁.E = Set.univ → M₂.E = Set.univ →      M₁.eRank = (r : ℕ∞) → M₂.eRank = (r : ℕ∞) →      ExactOracles M₁ M₂ O → 0 < ε → ε < 1 → 0 < δ → δ < 1 →      let B := resourceBound C d n r ε δ      let Z := commonBaseCount M₁ M₂      (∀ bits : ℕ → Bool,        let out := execution P O (initialConfig k s n r ε δ) bits B        out.output.isSome ∧ out.oracleCalls ≤ B ∧ out.bitOperations ≤ B ∧        ∀ z, out.output = some z → 0 ≤ z ∧ (Z = 0 → z = 0)) ∧      (1 - δ) * (2 : ℚ) ^ B ≤        (Nat.card {bits : Fin B → Bool //          ∃ z, (execution P O (initialConfig k s n r ε δ) (finiteBits bits) B).output            = some z ∧ relativeAccuracy ε Z z} : ℚ)theorem common_bases_fpras : CommonBasesFPRAS.MainTheorem := by  sorry

A Lean statement states the theorem and leaves its proof as sorry on purpose. OpenAI’s proof is in a separate solution module, and a program called Comparator checks that proof against this statement.

Definitions needed to read it

Local definitions the statement names, followed through the definitions they name in turn. The match is by name in the source text, so a name reached only through notation can be missed; the full file is below.

  • abbreviation OAI.CommonBasesFPRAS.IndependenceOracle line 7

    abbrev IndependenceOracle (n : ℕ) := Bool → (Fin n → Bool) → Bool
  • definition OAI.CommonBasesFPRAS.ExactOracles line 9

    def ExactOracles {n : ℕ} (M₁ M₂ : Matroid (Fin n))    (O : IndependenceOracle n) : Prop :=  ∀ (which : Bool) (A : Fin n → Bool),    O which A = true ↔ (if which then M₂ else M₁).Indep {i | A i = true}
  • definition OAI.CommonBasesFPRAS.commonBaseCount line 14

    noncomputable def commonBaseCount {n : ℕ} (M₁ M₂ : Matroid (Fin n)) : ℕ :=  Nat.card {A : Finset (Fin n) // M₁.IsBase (A : Set (Fin n)) ∧    M₂.IsBase (A : Set (Fin n))}
  • inductive type OAI.CommonBasesFPRAS.Instruction line 18

    inductive Instruction (k s : ℕ) where  | push (register : Fin (k + 8)) (bit : Bool) (next : Fin (s + 1))  | pop (register : Fin (k + 8)) (empty zero one : Fin (s + 1))  | coin (zero one : Fin (s + 1))  | query (which : Bool) (register : Fin (k + 8)) (no yes : Fin (s + 1))  | halt
  • definition OAI.CommonBasesFPRAS.Program line 25

    def Program (k s : ℕ) := Fin (s + 1) → Instruction k s
  • definition OAI.CommonBasesFPRAS.binaryValue line 27

    def binaryValue : List Bool → ℕ  | [] => 0  | b :: l => (if b then 1 else 0) + 2 * binaryValue l
  • structure OAI.CommonBasesFPRAS.MachineConfig line 31

    structure MachineConfig (k s : ℕ) where  registers : Fin (k + 8) → List Bool  control : Fin (s + 1)  output : Option ℚ := none  oracleCalls : ℕ := 0  bitOperations : ℕ := 0
  • definition OAI.CommonBasesFPRAS.inputLength line 38

    def inputLength (n r : ℕ) (ε δ : ℚ) : ℕ :=  6 + n.bits.length + r.bits.length + ε.num.natAbs.bits.length + ε.den.bits.length +    δ.num.natAbs.bits.length + δ.den.bits.length
  • definition OAI.CommonBasesFPRAS.initialConfig line 42

    def initialConfig (k s n r : ℕ) (ε δ : ℚ) : MachineConfig k s where  registers := fun i => match i.val with    | 0 => n.bits    | 1 => r.bits    | 2 => ε.num.natAbs.bits    | 3 => ε.den.bits    | 4 => δ.num.natAbs.bits    | 5 => δ.den.bits    | _ => []  control := 0
  • definition OAI.CommonBasesFPRAS.machineStep line 53

    def machineStep {k s n : ℕ} (P : Program k s) (O : IndependenceOracle n)    (randomBit : Bool) (c : MachineConfig k s) : MachineConfig k s :=
    Show the remaining 38 lines
      if c.output.isSome then c else  match P c.control with  | .push i b next =>    { c with      registers := Function.update c.registers i (b :: c.registers i)      control := next      bitOperations := c.bitOperations + 1 }  | .pop i empty zero one =>    match c.registers i with    | [] => { c with      control := empty      bitOperations := c.bitOperations + 1 }    | b :: l =>      { c with      registers := Function.update c.registers i l      control := if b then one else zero      bitOperations := c.bitOperations + 1 }  | .coin zero one =>    { c with      control := if randomBit then one else zero      bitOperations := c.bitOperations + 1 }  | .query which i no yes =>    if (c.registers i).length = n then      { c with      control := if O which (fun j => (c.registers i)[j.val]?.getD false)          then yes else no      oracleCalls := c.oracleCalls + 1      bitOperations := c.bitOperations + (c.registers i).length + n + 1 }    else      { c with      output := some 0      bitOperations := c.bitOperations + (c.registers i).length + n + 1 }  | .halt =>    { c with      output := some ((binaryValue (c.registers ⟨6, by omega⟩) : ℚ) /        (binaryValue (c.registers ⟨7, by omega⟩) : ℚ))      bitOperations := c.bitOperations + (c.registers ⟨6, by omega⟩).length +        (c.registers ⟨7, by omega⟩).length + 1 }
  • definition OAI.CommonBasesFPRAS.execution line 94

    def execution {k s n : ℕ} (P : Program k s) (O : IndependenceOracle n)    (c : MachineConfig k s) (bits : ℕ → Bool) : ℕ → MachineConfig k s  | 0 => c  | t + 1 => machineStep P O (bits t) (execution P O c bits t)
  • definition OAI.CommonBasesFPRAS.finiteBits line 99

    def finiteBits {B : ℕ} (bits : Fin B → Bool) (t : ℕ) : Bool :=  if h : t < B then bits ⟨t, h⟩ else false
  • definition OAI.CommonBasesFPRAS.relativeAccuracy line 102

    def relativeAccuracy (ε : ℚ) (Z : ℕ) (z : ℚ) : Prop :=  (1-ε)*(Z:ℚ)≤z ∧ z≤(1+ε)*(Z:ℚ)
  • definition OAI.CommonBasesFPRAS.resourceBound line 105

    def resourceBound (C d n r : ℕ) (ε δ : ℚ) : ℕ :=  C * (1 + n + inputLength n r ε δ + ⌈ε⁻¹⌉₊ + Nat.clog 2 ⌈δ⁻¹⌉₊) ^ d

Possible paper match, with no recorded review

Atlas selected this possible match by comparing words in the Lean statement with every statement in the family. No review of whether the statements mean the same thing is recorded.

Words of the Lean statement, its names split into words, that also appear in the theorem: independence, zero, calls, matroid, oracle, random, output, binary and seven more. Found in the paper’s title, which its main theorem states: bases, counting and common. Shared only with the release’s name for the Lean statement or OpenAI’s note on it: prob, delta, randomized, number, nonnegative and polynomial.

Statements in Approximate counting of common bases of two matroids come first because the release lists only that paper for this Lean statement.

Theorem 1.1 of Approximate counting of common bases of two matroids

Theorem

There is a single randomized oracle algorithm with the following guarantee. Given independence oracles for arbitrary rank-r matroids M_1,M_2 on [n], and rational \epsilon,\delta\in(0,1), it outputs a nonnegative rational \widehat Z such that

\Prob\bigl[(1-\epsilon)Z(M_1,M_2)\le\widehat Z \le(1+\epsilon)Z(M_1,M_2)\bigr]\ge1-\delta.

If Z(M_1,M_2)=0, its output is always zero. On every execution, the number of oracle calls and the number of other bit operations are bounded by a fixed polynomial in

n,\qquad \ell,\qquad \epsilon^{-1},\qquad \log\delta^{-1},

where \ell is the binary encoding length of n,r,\epsilon,\delta. The algorithm uses only independent unbiased random bits and does not require a representation of either matroid.

Section 1, Introduction Read it in the paper TeX source, line 86
Show the TeX
There is a single randomized oracle algorithm with the following
guarantee.  Given independence oracles for arbitrary rank-$r$ matroids
$M_1,M_2$ on $[n]$, and rational $\epsilon,\delta\in(0,1)$, it outputs a
nonnegative rational $\widehat Z$ such that
\[
 \Prob\bigl[(1-\epsilon)Z(M_1,M_2)\le\widehat Z
                  \le(1+\epsilon)Z(M_1,M_2)\bigr]\ge1-\delta.
\]
If $Z(M_1,M_2)=0$, its output is always zero.  On every execution, the
number of oracle calls and the number of other bit operations are bounded
by a fixed polynomial in
\[
 n,\qquad \ell,\qquad \epsilon^{-1},\qquad \log\delta^{-1},
\]
where $\ell$ is the binary encoding length of $n,r,\epsilon,\delta$.
The algorithm uses only independent unbiased random bits and does not
require a representation of either matroid.

Other possible paper matches, closest wording first

  • Corollary 9.1 of Approximate counting of common bases of two matroids. Words of the Lean statement, its names split into words, that also appear in the theorem: independence, calls, queries, matroids, oracle, common and twelve more. Shared only with the release’s name for the Lean statement or OpenAI’s note on it: enumerated, two, ground, prob, delta, randomized and two more.
  • Theorem 1.1 of An FPRAS for Common Integer Polymatroid Bases with Binary Capacities. Words of the Lean statement, its names split into words, that also appear in the theorem: zero, one, calls, oracle, output, execution and six more. Found in the paper’s title, which its main theorem states: fpras, bases, binary, common and integer. Shared only with the release’s name for the Lean statement or OpenAI’s note on it: delta, randomized, number, nonnegative and polynomial.
  • Proposition 2.1 of An FPRAS for Common Integer Polymatroid Bases with Binary Capacities. Words of the Lean statement, its names split into words, that also appear in the theorem: value, finite, query, function, oracle, binary and seven more. Shared only with the release’s name for the Lean statement or OpenAI’s note on it: ground and polynomial.
  • Proposition 6.3 of An FPRAS for Common Integer Polymatroid Bases with Binary Capacities. Words of the Lean statement, its names split into words, that also appear in the theorem: queries, random, output, rank, execution, length and five more. Shared only with the release’s name for the Lean statement or OpenAI’s note on it: using, approximation, number and polynomial.
  • Proposition 4.5 of An FPRAS for Common Integer Polymatroid Bases with Binary Capacities. Words of the Lean statement, its names split into words, that also appear in the theorem: base, count and common.
  • Proposition 7.2 of An FPRAS for Common Integer Polymatroid Bases with Binary Capacities. Words of the Lean statement, its names split into words, that also appear in the theorem: step, initial, function, one, counting and input.

How Atlas selects possible matches: it compares words in the Lean statement with every statement in the family. It also splits Lean names into words. The release’s label and note receive less weight; a main theorem wins a close match. If the Lean wording yields no clear match, Atlas uses the catalogue description. This selects passages to compare; it does not establish that they mean the same thing.

Reported Comparator runs

three independent groups report that Comparator accepted runs for this Lean statement file: Carlos Toledo’s cert-machine; MathVet; Neal McBurnett’s openai-math-lean-tooling. Each group reports its own Comparator run for this statement file. Atlas has not repeated the runs.

For a configured theorem, Comparator tests whether OpenAI’s proof proves the exact Lean statement. Some files instead configure definitions for comparison by name and type. Acceptance does not establish that the statement means the same thing as the paper’s theorem.

Carlos Toledo’s cert-machine

How they ran it, in their words

Every Comparator challenge re-run on Linux with nanoda switched on, decided by a preregistered rule; every family’s challenge statements read against OpenAI’s headline, clause by clause. Decider: Comparator (leanprover/comparator d03acab1 = tag v4.34.0, built with the release’s toolchain v4.34.1), lean4export 076e8e57, landrun 811cfff5, nanoda_lib 3a240721 — on GitHub’s Linux runners (.github/workflows/openai-math-kernel.yml), one j…

  • Accepted Their term: “CERTIFIED”.

    Re-run through Comparator with the independent nanoda kernel switched on: Lean’s kernel and nanoda both accept OpenAI’s solution.

    Reported 10 October 2026, at the release version shown here. Original verification report

MathVet

How they ran it, in their words

An independent, model-assisted, pre-referee reading of each family’s challenge statements against its headline claim under a published rubric (PROTOCOL.md), with its own Comparator runs per challenge. Its protocol describes a planned error-rate study; this import does not establish that the study has been completed.

  • Accepted Their term: “comparator-sandboxed-pass”.

    Re-run through Comparator in a sandboxed virtual machine: accepted.

    Reported 8 October 2026, at the release version shown here. Original verification report

Neal McBurnett’s openai-math-lean-tooling

How they ran it, in their words

Each challenge’s solution is built and checked by Comparator at the pinned release, on Lean leanprover/lean4:v4.34.1, recording the verdict, the check time and the peak memory; the run was still in progress at this snapshot (192 accepted, 224 pending).

  • Accepted Their term: “accepted”.

    Re-run through Comparator with Lean’s kernel: accepted (the check took 79 s).

    Reported 10 October 2026, at the release version shown here. Original verification report

What an accepted Comparator run would establish

If Comparator is run under this configuration and accepts, the solution proves the configured theorem using only the allowed axioms, and the Lean kernel accepts every step.

This configuration lists no definition holes, so every definition the statement uses is compared as it is written in this file.

Allowed axioms: propext, Quot.sound, Classical.choice

Acceptance leaves open whether the Lean statement means the same thing as the paper’s theorem, whether the paper’s other claims are correct, and whether the result is new. Those questions require comparing the definitions and statements, reviewing the paper’s arguments, and reading the earlier literature.

Do family 114’s Lean statements express the advertised result?

These assessments compare all the family’s Lean statements with the whole result OpenAI describes in its catalogue. They do not assess this statement alone. Each group uses its own published rules.

Neither group finds that the Lean statements express or imply the whole advertised result. They describe what the statements express in different terms.

  • MathVet

    Finds part of it stated, or a narrower or weaker statement Their term: “partial”.

    Only a special case is stated, or a co-equal headline claim has no challenge statement. Their record

    Their note: Matroid special case only (independence oracles, equal ranks); the integer-polymatroid FPRAS with binary capacities and rank-value oracles, and the variants for common independent sets of other cardinalities/unequal ranks, are not stated.

    Definitions the group asks readers to compare with the paper: Hand-rolled oracle register machine: Instruction = push/pop/coin/query/halt over Fin (k+8) bit-list registers with counters oracleCalls/bitOperations; a query on a register of length ≠ n silently outputs 0 and stops; ExactOracles says the oracle answers independence of the characteristic vector exactly; inputLength counts n and r in binary while the resource bound is polynomial in n itself

  • Carlos Toledo’s cert-machine

    Finds part of it stated, or a narrower or weaker statement Their term: “NARROWER — UNDECLARED”.

    The challenge statements state less than the headline, and OpenAI’s Lean note does not say so. Their record

    Their note: The headline object (common integer bases of integral polymatroids with binary capacities and rank-value oracles) and the independent-set variants are not stated. Only the matroid common-bases FPRAS is stated (the September 23 paper; the October 5 polymatroid paper is not formalized).

Read the complete assessments on the family page.

OpenAI’s own note on what this Lean statement covers

“The formalized result gives a fully polynomial randomized approximation scheme for the number of common bases of two equal-rank matroids on an enumerated finite ground set, using independence oracles. For rational $0<\varepsilon,\delta<1$, the nonnegative rational output has relative error at most $\varepsilon$ with probability at least $1-\delta$.”

An AI agent drafted an interpretation of that note, classifying this statement as “states a paper’s theorem as the paper does”. No human review of the draft is recorded.

Supporting material

The complete Lean statement file

131 lines, copied from release version fd4aeeb. A name defined in this file links to its line; pointing at the name shows its definition.

Declared in this fileOpenAI’s library at fd4aeebOther linked names open the current documentation for Mathlib, Lean’s mathematical library.

Links in the file: 29 to its own declarations, 60 to Mathlib’s documentation. 16 declared names also link to the same declaration in OpenAI’s library. OpenAI uses Mathlib version d13f23b. The linked documentation describes the current version, so names that have since changed may be missing there.

import Mathlibnamespace OAInamespace CommonBasesFPRASabbrev IndependenceOracle (n : ℕ) := Bool → (Fin n → Bool) → Booldef ExactOracles {n : ℕ} (M₁ M₂ : Matroid (Fin n))    (O : IndependenceOracle n) : Prop :=  ∀ (which : Bool) (A : Fin n → Bool),    O which A = true ↔ (if which then M₂ else M₁).Indep {i | A i = true}noncomputable def commonBaseCount {n : ℕ} (M₁ M₂ : Matroid (Fin n)) : ℕ :=  Nat.card {A : Finset (Fin n) // M₁.IsBase (A : Set (Fin n)) ∧    M₂.IsBase (A : Set (Fin n))}inductive Instruction (k s : ℕ) where  | push (register : Fin (k + 8)) (bit : Bool) (next : Fin (s + 1))  | pop (register : Fin (k + 8)) (empty zero one : Fin (s + 1))  | coin (zero one : Fin (s + 1))  | query (which : Bool) (register : Fin (k + 8)) (no yes : Fin (s + 1))  | haltdef Program (k s : ℕ) := Fin (s + 1) → Instruction k sdef binaryValue : List Bool → ℕ  | [] => 0  | b :: l => (if b then 1 else 0) + 2 * binaryValue lstructure MachineConfig (k s : ℕ) where  registers : Fin (k + 8) → List Bool  control : Fin (s + 1)  output : Option ℚ := none  oracleCalls : ℕ := 0  bitOperations : ℕ := 0def inputLength (n r : ℕ) (ε δ : ℚ) : ℕ :=  6 + n.bits.length + r.bits.length + ε.num.natAbs.bits.length + ε.den.bits.length +    δ.num.natAbs.bits.length + δ.den.bits.lengthdef initialConfig (k s n r : ℕ) (ε δ : ℚ) : MachineConfig k s where  registers := fun i => match i.val with    | 0 => n.bits    | 1 => r.bits    | 2 => ε.num.natAbs.bits    | 3 => ε.den.bits    | 4 => δ.num.natAbs.bits    | 5 => δ.den.bits    | _ => []  control := 0def machineStep {k s n : ℕ} (P : Program k s) (O : IndependenceOracle n)    (randomBit : Bool) (c : MachineConfig k s) : MachineConfig k s :=  if c.output.isSome then c else  match P c.control with  | .push i b next =>    { c with      registers := Function.update c.registers i (b :: c.registers i)      control := next      bitOperations := c.bitOperations + 1 }  | .pop i empty zero one =>    match c.registers i with    | [] => { c with      control := empty      bitOperations := c.bitOperations + 1 }    | b :: l =>      { c with      registers := Function.update c.registers i l      control := if b then one else zero      bitOperations := c.bitOperations + 1 }  | .coin zero one =>    { c with      control := if randomBit then one else zero      bitOperations := c.bitOperations + 1 }  | .query which i no yes =>    if (c.registers i).length = n then      { c with      control := if O which (fun j => (c.registers i)[j.val]?.getD false)          then yes else no      oracleCalls := c.oracleCalls + 1      bitOperations := c.bitOperations + (c.registers i).length + n + 1 }    else      { c with      output := some 0      bitOperations := c.bitOperations + (c.registers i).length + n + 1 }  | .halt =>    { c with      output := some ((binaryValue (c.registers ⟨6, by omega⟩) : ℚ) /        (binaryValue (c.registers ⟨7, by omega⟩) : ℚ))      bitOperations := c.bitOperations + (c.registers ⟨6, by omega⟩).length +        (c.registers ⟨7, by omega⟩).length + 1 }def execution {k s n : ℕ} (P : Program k s) (O : IndependenceOracle n)    (c : MachineConfig k s) (bits : ℕ → Bool) : ℕ → MachineConfig k s  | 0 => c  | t + 1 => machineStep P O (bits t) (execution P O c bits t)def finiteBits {B : ℕ} (bits : Fin B → Bool) (t : ℕ) : Bool :=  if h : t < B then bits ⟨t, h⟩ else falsedef relativeAccuracy (ε : ℚ) (Z : ℕ) (z : ℚ) : Prop :=  (1-ε)*(Z:ℚ)≤z ∧ z≤(1+ε)*(Z:ℚ)def resourceBound (C d n r : ℕ) (ε δ : ℚ) : ℕ :=  C * (1 + n + inputLength n r ε δ + ⌈ε⁻¹⌉₊ + Nat.clog 2 ⌈δ⁻¹⌉₊) ^ ddef MainTheorem : Prop :=  ∃ (k s : ℕ) (P : Program k s) (C d : ℕ), 0 < C ∧    ∀ (n r : ℕ) (M₁ M₂ : Matroid (Fin n)) (O : IndependenceOracle n)      (ε δ : ℚ),      M₁.E = Set.univ → M₂.E = Set.univ →      M₁.eRank = (r : ℕ∞) → M₂.eRank = (r : ℕ∞) →      ExactOracles M₁ M₂ O → 0 < ε → ε < 1 → 0 < δ → δ < 1 →      let B := resourceBound C d n r ε δ      let Z := commonBaseCount M₁ M₂      (∀ bits : ℕ → Bool,        let out := execution P O (initialConfig k s n r ε δ) bits B        out.output.isSome ∧ out.oracleCalls ≤ B ∧ out.bitOperations ≤ B ∧        ∀ z, out.output = some z → 0 ≤ z ∧ (Z = 0 → z = 0)) ∧      (1 - δ) * (2 : ℚ) ^ B ≤        (Nat.card {bits : Fin B → Bool //          ∃ z, (execution P O (initialConfig k s n r ε δ) (finiteBits bits) B).output            = some z ∧ relativeAccuracy ε Z z} : ℚ)end CommonBasesFPRAStheorem common_bases_fpras : CommonBasesFPRAS.MainTheorem := by  sorryend OAI
Every extracted statement of the accompanying papers

The statements the atlas extracted automatically from each accompanying paper’s TeX source, with links to the full sources.

Approximate counting of common bases of two matroids

Theorem

There is a single randomized oracle algorithm with the following guarantee. Given independence oracles for arbitrary rank-r matroids M_1,M_2 on [n], and rational \epsilon,\delta\in(0,1), it outputs a nonnegative rational \widehat Z such that

\Prob\bigl[(1-\epsilon)Z(M_1,M_2)\le\widehat Z \le(1+\epsilon)Z(M_1,M_2)\bigr]\ge1-\delta.

If Z(M_1,M_2)=0, its output is always zero. On every execution, the number of oracle calls and the number of other bit operations are bounded by a fixed polynomial in

n,\qquad \ell,\qquad \epsilon^{-1},\qquad \log\delta^{-1},

where \ell is the binary encoding length of n,r,\epsilon,\delta. The algorithm uses only independent unbiased random bits and does not require a representation of either matroid.

Section 1, Introduction Read it in the paper TeX source, line 86
Show the TeX
There is a single randomized oracle algorithm with the following
guarantee.  Given independence oracles for arbitrary rank-$r$ matroids
$M_1,M_2$ on $[n]$, and rational $\epsilon,\delta\in(0,1)$, it outputs a
nonnegative rational $\widehat Z$ such that
\[
 \Prob\bigl[(1-\epsilon)Z(M_1,M_2)\le\widehat Z
                  \le(1+\epsilon)Z(M_1,M_2)\bigr]\ge1-\delta.
\]
If $Z(M_1,M_2)=0$, its output is always zero.  On every execution, the
number of oracle calls and the number of other bit operations are bounded
by a fixed polynomial in
\[
 n,\qquad \ell,\qquad \epsilon^{-1},\qquad \log\delta^{-1},
\]
where $\ell$ is the binary encoding length of $n,r,\epsilon,\delta$.
The algorithm uses only independent unbiased random bits and does not
require a representation of either matroid.
Configuration and file facts
OpenAI’s Lean catalogue
OpenAI lists this as a main result in its Lean catalogue.
Allowed axioms
propext, Quot.sound, Classical.choiceThe axioms this configuration allows a proof to use. A replay shows which of them a proof uses.
Imports
Mathlib
Imports from outside Mathlib
None
Local vocabulary
15 local vocabulary declarations
Lines of Lean, without comments or blank lines
106
Second proof-verification program (nanoda)
Off in OpenAI's configuration.

Statement comparison

A shareable summary of this page: the proposed paper match, what the Lean statement expresses, and which reviews are recorded.

Plectis Atlas

FPRAS for common matroid bases

Family 114: Approximate counting of common integer polymatroid bases

Candidate pairing

Theorem

There is a single randomized oracle algorithm with the following guarantee. Given independence oracles for arbitrary rank-r matroids M_1,M_2 on [n], and rational \epsilon,\delta\in(0,1), it outputs a nonnegative rational \widehat Z such that

\Prob\bigl[(1-\epsilon)Z(M_1,M_2)\le\widehat Z \le(1+\epsilon)Z(M_1,M_2)\bigr]\ge1-\delta.

If Z(M_1,M_2)=0, its output is always zero. On every execution, the number of oracle calls and the number of other bit operations are bounded by a fixed polynomial in

n,\qquad \ell,\qquad \epsilon^{-1},\qquad \log\delta^{-1},

where \ell is the binary encoding length of n,r,\epsilon,\delta. The algorithm uses only independent unbiased random bits and does not require a representation of either matroid.

Approximate counting of common bases of two matroids

TeX label thm:main, build/main.tex, line 86

The Lean statement says

def MainTheorem : Prop :=  ∃ (k s : ℕ) (P : Program k s) (C d : ℕ), 0 < C ∧    ∀ (n r : ℕ) (M₁ M₂ : Matroid (Fin n)) (O : IndependenceOracle n)      (ε δ : ℚ),      M₁.E = Set.univ → M₂.E = Set.univ →      M₁.eRank = (r : ℕ∞) → M₂.eRank = (r : ℕ∞) →      ExactOracles M₁ M₂ O → 0 < ε → ε < 1 → 0 < δ → δ < 1 →      let B := resourceBound C d n r ε δ      let Z := commonBaseCount M₁ M₂      (∀ bits : ℕ → Bool,        let out := execution P O (initialConfig k s n r ε δ) bits B        out.output.isSome ∧ out.oracleCalls ≤ B ∧ out.bitOperations ≤ B ∧        ∀ z, out.output = some z → 0 ≤ z ∧ (Z = 0 → z = 0)) ∧      (1 - δ) * (2 : ℚ) ^ B ≤        (Nat.card {bits : Fin B → Bool //          ∃ z, (execution P O (initialConfig k s n r ε δ) (finiteBits bits) B).output            = some z ∧ relativeAccuracy ε Z z} : ℚ)theorem common_bases_fpras : CommonBasesFPRAS.MainTheorem := by

A Lean statement states the theorem and leaves its proof as sorry on purpose. OpenAI’s proof is in a separate solution module, and a program called Comparator checks that proof against this statement.

Atlas selected this possible match because the two statements share words. No review of whether they mean the same thing is recorded.three independent groups report that Comparator accepted runs for this file at the release version shown here.Statement review: none recorded

15 local vocabulary declarationsAllowed axioms: propext, Quot.sound, Classical.choiceopenai/math at fd4aeebwcook04.github.io/plectis-atlas/statements/CommonBasesFPRAS.html

Recorded comparisons with the paper

Atlas records no contributed comparison of this statement with the paper.

Contributed Comparator run reports

Atlas records no contributed report of running Comparator for this statement file.

Questions to resolve when comparing this statement with the paper

  • Both independent readers find this family's challenges state less than its headline, while the atlas's AI draft classes this file as stating a paper's theorem as the paper does; check which part of the headline it covers.

Compare this statement with the paperShare a Comparator run reportSuggest a change