Plectis Atlas
Search

↑ and ↓ move, Enter opens, Esc closes.

Lean statement

Unconditional sequential time lower bounds for Weisfeiler–Leman equivalence

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.WLTime.unconditional_time_lower_bound

Lean statement file: WeisfeilerLeman.leanShow on the mapOpenAI’s Lean catalogue does not list this as a main result; the release’s Lean notes for family 133 link it.

The Lean statement

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

def MainClaim : Prop :=  ∃ c : ℝ, 0 < c ∧ ∃ k₀ : ℕ, ∀ k : ℕ, k₀ ≤ k →    ∀ convention : Convention, ∀ p : InputClass, ∀ A : Model,      Decides A convention k p → ∃ n₀ : ℕ, ∀ n : ℕ, n₀ ≤ n →        (n : ℝ) ^ (c * (k : ℝ)) ≤ (worstTime A p n : ℝ)theorem WLTime.unconditional_time_lower_bound : WLTime.MainClaim := 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.

  • inductive type OAI.WLTime.Convention line 7

    inductive Convention where  | joint | separate  deriving DecidableEq
  • definition OAI.WLTime.Graph line 11

    def Graph (n : ℕ) :=  { a : Fin n → Fin n → Bool //    (∀ x y, a x y = a y x) ∧ (∀ x, a x x = false) }
  • definition OAI.WLTime.Graph.adj line 17

    def adj {n : ℕ} (G : Graph n) : Fin n → Fin n → Bool := G.val
  • definition OAI.WLTime.Graph.diameterTwo line 19

    def diameterTwo {n : ℕ} (G : Graph n) : Prop :=  0 < n ∧ ∀ x y, x = y ∨ G.adj x y = true ∨    ∃ z, G.adj x z = true ∧ G.adj z y = true
  • abbreviation OAI.WLTime.AtomicColor line 30

    abbrev AtomicColor (k : ℕ) :=  (Fin k → Fin k → Bool) × (Fin k → Fin k → Bool)
  • definition OAI.WLTime.Color line 33

    def Color (c : Convention) (k : ℕ) : ℕ → Type  | 0 => AtomicColor k  | t + 1 => match c with    | .joint => Color c k t × Multiset (Fin k → Color c k t)    | .separate => Color c k t × (Fin k → Multiset (Color c k t))
  • definition OAI.WLTime.tupleColor line 39

    noncomputable def tupleColor {n : ℕ} (c : Convention) (k : ℕ) (G : Graph n) :    (t : ℕ) → (Fin k → Fin n) → Color c k t  | 0, v => (fun i j => decide (v i = v j), fun i j => G.adj (v i) (v j))  | t + 1, v => match c with    | .joint => (tupleColor .joint k G t v,        Finset.univ.val.map (fun z => fun i =>          tupleColor .joint k G t (Function.update v i z)))    | .separate => (tupleColor .separate k G t v,        fun i => Finset.univ.val.map (fun z =>          tupleColor .separate k G t (Function.update v i z)))
  • definition OAI.WLTime.histogram line 50

    noncomputable def histogram {n : ℕ} (c : Convention) (k : ℕ) (G : Graph n)    (t : ℕ) : Multiset (Color c k t) :=  Finset.univ.val.map (tupleColor c k G t)
  • definition OAI.WLTime.Equivalent line 54

    def Equivalent {n : ℕ} (c : Convention) (k : ℕ) (G H : Graph n) : Prop :=  ∀ t, histogram c k G t = histogram c k H t
  • definition OAI.WLTime.encodeNat line 57

    def encodeNat (n : ℕ) : List Bool :=  let bits := (Nat.digits 2 n).map (fun b => decide (b = 1))  List.replicate bits.length true ++ [false] ++ bits
  • definition OAI.WLTime.matrixBits line 61

    def matrixBits {n : ℕ} (G : Graph n) : List Bool :=  (List.ofFn fun i : Fin n => List.ofFn fun j : Fin n => G.adj i j).flatten
  • definition OAI.WLTime.encodePair line 64

    def encodePair {n : ℕ} (G H : Graph n) : List Bool :=  encodeNat n ++ matrixBits G ++ matrixBits H
  • inductive type OAI.WLTime.Move line 67

    inductive Move where  | left | stay | right  deriving DecidableEq
  • definition OAI.WLTime.Move.displacement line 71

    def Move.displacement : Move → ℤ  | .left => -1  | .stay => 0  | .right => 1
  • structure OAI.WLTime.TM line 76

    structure TM where  tapes : ℕ  states : ℕ  symbols : ℕ  tapes_pos : 0 < tapes  states_ge : 3 ≤ states  symbols_ge : 3 ≤ symbols  transition : Fin states → (Fin tapes → Fin symbols) →    Fin states × (Fin tapes → Fin symbols) × (Fin tapes → Move)
  • structure OAI.WLTime.TM.Config line 88

    structure Config (M : TM) where  state : Fin M.states  head : Fin M.tapes → ℤ  tape : Fin M.tapes → ℤ → Fin M.symbols
  • definition OAI.WLTime.TM.symbol line 93

    def symbol (M : TM) (b : Bool) : Fin M.symbols :=  if b then ⟨2, by have := M.symbols_ge; omega⟩ else    ⟨1, by have := M.symbols_ge; omega⟩
  • definition OAI.WLTime.TM.initial line 97

    def initial (M : TM) (w : List Bool) : M.Config where  state := ⟨0, by have := M.states_ge; omega⟩  head := fun _ => 0  tape := fun a z => if a.val = 0 ∧ 0 ≤ z ∧ z.toNat < w.length then    M.symbol (w.getD z.toNat false) else ⟨0, by have := M.symbols_ge; omega⟩
  • definition OAI.WLTime.TM.halted line 103

    def halted (M : TM) (C : M.Config) : Prop := C.state.val = 1 ∨ C.state.val = 2
  • definition OAI.WLTime.TM.step line 108

    def step (M : TM) (C : M.Config) : M.Config :=  if M.halted C then C else  let tr := M.transition C.state (fun a => C.tape a (C.head a))  { state := tr.1    head := fun a => C.head a + (tr.2.2 a).displacement    tape := fun a z => if z = C.head a then tr.2.1 a else C.tape a z }
  • definition OAI.WLTime.TM.run line 115

    def run (M : TM) (w : List Bool) (t : ℕ) : M.Config := M.step^[t] (M.initial w)
  • definition OAI.WLTime.TM.halts line 117

    def halts (M : TM) (w : List Bool) : Prop := ∃ t, M.halted (M.run w t)
  • definition OAI.WLTime.TM.time line 119

    noncomputable def time (M : TM) (w : List Bool) : ℕ := by  classical  exact if h : M.halts w then Nat.find h else 0
  • definition OAI.WLTime.TM.accepts line 123

    def accepts (M : TM) (w : List Bool) : Prop :=  ∃ t, (M.run w t).state.val = 1

13 more definitions are reached the same way; the full file lists them.

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: instruction, multiset, memory, log, diameter, applies, separate, pairs and 14 more. Found in the paper’s title, which its main theorem states: unconditional and lower. Shared only with the release’s name for the Lean statement or OpenAI’s note on it: required, sequential, ck, dimensional, absolute, leman, multitape, weisfeiler and two more.

Statements in Unconditional time lower bounds for Weisfeiler–Leman equivalence come first because the release’s name for this Lean statement shares words with that paper’s title.

Theorem 1.1 of Unconditional time lower bounds for Weisfeiler–Leman equivalence

Theorem

There are absolute constants c>0 and k_0 with the following property. For every fixed integer k\ge k_0 and every correct deterministic decider A for k-Weisfeiler–Leman equivalence, there is an integer n_0=n_0(A,k) such that

T_A(n)\ge n^{ck} \qquad\text{for every integer }n\ge n_0.

This holds for multitape Turing machines and for sequential random-access machines with O(\log(n+2))-bit addresses and words, boundedly many memory accesses per instruction, and fixed instructions computable in time polynomial in the word length. The same bound holds if A is required to decide correctly only on pairs of simple connected uncolored graphs of diameter at most two, with T_A(n) maximized over that class. It applies both to the joint replacement-multiset convention for k-dimensional refinement and to the separate-multiset convention defined below.

Section 1, Introduction Read it in the paper TeX source, line 82
Show the TeX
There are absolute constants $c>0$ and $k_0$ with the following
property. For every fixed integer $k\ge k_0$ and every correct
deterministic decider $A$ for $k$-Weisfeiler--Leman equivalence, there
is an integer $n_0=n_0(A,k)$ such that
\[
                     T_A(n)\ge n^{ck}
             \qquad\text{for every integer }n\ge n_0.
\]
This holds for multitape Turing machines and for sequential
random-access machines with $O(\log(n+2))$-bit addresses and words,
boundedly many memory accesses per instruction, and fixed instructions
computable in time polynomial in the word length.
The same bound holds if $A$ is required to decide correctly only on
pairs of simple connected uncolored graphs of diameter at most two,
with $T_A(n)$ maximized over that class. It applies both to the
joint replacement-multiset convention for $k$-dimensional refinement
and to the separate-multiset convention defined below.

Other possible paper matches, closest wording first

  • Theorem 4.3 of Unconditional time lower bounds for Weisfeiler–Leman equivalence. Words of the Lean statement, its names split into words, that also appear in the theorem: halts, update, joint, wl, input, accepts and eight more. Shared only with the release’s name for the Lean statement or OpenAI’s note on it: sufficiently, dimensional, absolute, multitape, large, order and one more.
  • Theorem 5.2 of Unconditional time lower bounds for Weisfeiler–Leman equivalence. Words of the Lean statement, its names split into words, that also appear in the theorem: worst, tape, update, accepting, input, constant and four more. Shared only with the release’s name for the Lean statement or OpenAI’s note on it: case, sufficiently, large, deterministic and one.
  • Proposition 2.3 of Variable-dimension Weisfeiler–Leman equivalence on general and subcubic graphs. Words of the Lean statement, its names split into words, that also appear in the theorem: exponents, log, decided, bit, wl, constants and five more. Shared only with the release’s name for the Lean statement or OpenAI’s note on it: absolute and order.
  • Theorem 1.3 of Parity lifts and bounded-treewidth witnesses for Weisfeiler–Leman equivalence. Words of the Lean statement, its names split into words, that also appear in the theorem: decides, separate, turing, joint, constants, inputs and seven more. Shared only with the release’s name for the Lean statement or OpenAI’s note on it: ck, algorithm, multitape, equivalence and deterministic.
  • Theorem 3.2 of Unconditional time lower bounds for Weisfeiler–Leman equivalence. Words of the Lean statement, its names split into words, that also appear in the theorem: diameter, pair, bound, two and graphs. Shared only with the release’s name for the Lean statement or OpenAI’s note on it: large, equivalence, order and one.
  • Theorem 1.2 of Parity lifts and bounded-treewidth witnesses for Weisfeiler–Leman equivalence. Words of the Lean statement, its names split into words, that also appear in the theorem: histograms, tuple, color, eq, separate, convention and three more. Shared only with the release’s name for the Lean statement or OpenAI’s note on it: order and one.

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 238 s; peak memory 6.2 GiB).

    Reported 11 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 133’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.

MathVet and cert-machine agree on how much of the advertised result the Lean statements express.

  • MathVet

    Finds the whole advertised result stated or implied Their term: “full”.

    The challenge statements imply the headline claim after unpacking definitions or one routine step. Their record

    Their note: None; all three headline claims are stated. The conditional running-time exclusions are not (docs).

    Definitions the group asks readers to compare with the paper: Hand-rolled computation models: TM (multi-tape, fixed finite tables, halting states 1/2), a RAM with logarithmic word width (WordWidth.logarithmic), memory Word w → Word w, compute (op : WordOperation R) where a whole word operation (any map implemented by a TM within constant * (w+1)^exponent) counts as one instruction and one step runs a finite instruction tree; InEXPTIME (2^(c (n+1)^d) steps), PolytimeReduces/PolytimeManyOne, EXPTIMEComplete; WL colors are canonical nested multisets (Color c k t) with Equivalent := ∀ t, histogram G t = histogram H t

  • Carlos Toledo’s cert-machine

    Finds the whole advertised result stated or implied Their term: “MATCHES”.

    The challenge statements state the headline as OpenAI wrote it. Their record

Read the complete assessments on the family page.

OpenAI’s own note on what this Lean statement covers

“The formalization proves an unconditional deterministic time lower bound for deciding $k$-dimensional Weisfeiler–Leman equivalence. For every sufficiently large fixed $k$, every deciding algorithm in the stated multitape Turing-machine or sequential logarithmic-word RAM model requires at least $n^{ck}$ worst-case time for all sufficiently large graph orders $n$, for one absolute $c>0$.”

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

270 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: 104 to its own declarations, 134 to Mathlib’s documentation. 51 declared names also link to the same declaration in OpenAI’s library. 1 identifier stays plain because it resolves to no single declaration: a variable Lean binds automatically, notation, or a name the scan of the pinned sources did not find. 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 WLTimeinductive Convention where  | joint | separate  deriving DecidableEqdef Graph (n : ℕ) :=  { a : Fin n → Fin n → Bool //    (∀ x y, a x y = a y x) ∧ (∀ x, a x x = false) }namespace Graphdef adj {n : ℕ} (G : Graph n) : Fin n → Fin n → Bool := G.valdef diameterTwo {n : ℕ} (G : Graph n) : Prop :=  0 < n ∧ ∀ x y, x = y ∨ G.adj x y = true ∨    ∃ z, G.adj x z = true ∧ G.adj z y = truenoncomputable instance (n : ℕ) : Fintype (Graph n) := by  classical  unfold Graph  infer_instanceend Graphabbrev AtomicColor (k : ℕ) :=  (Fin k → Fin k → Bool) × (Fin k → Fin k → Bool)def Color (c : Convention) (k : ℕ) : ℕ → Type  | 0 => AtomicColor k  | t + 1 => match c with    | .joint => Color c k t × Multiset (Fin k → Color c k t)    | .separate => Color c k t × (Fin k → Multiset (Color c k t))noncomputable def tupleColor {n : ℕ} (c : Convention) (k : ℕ) (G : Graph n) :    (t : ℕ) → (Fin k → Fin n) → Color c k t  | 0, v => (fun i j => decide (v i = v j), fun i j => G.adj (v i) (v j))  | t + 1, v => match c with    | .joint => (tupleColor .joint k G t v,        Finset.univ.val.map (fun z => fun i =>          tupleColor .joint k G t (Function.update v i z)))    | .separate => (tupleColor .separate k G t v,        fun i => Finset.univ.val.map (fun z =>          tupleColor .separate k G t (Function.update v i z)))noncomputable def histogram {n : ℕ} (c : Convention) (k : ℕ) (G : Graph n)    (t : ℕ) : Multiset (Color c k t) :=  Finset.univ.val.map (tupleColor c k G t)def Equivalent {n : ℕ} (c : Convention) (k : ℕ) (G H : Graph n) : Prop :=  ∀ t, histogram c k G t = histogram c k H tdef encodeNat (n : ℕ) : List Bool :=  let bits := (Nat.digits 2 n).map (fun b => decide (b = 1))  List.replicate bits.length true ++ [false] ++ bitsdef matrixBits {n : ℕ} (G : Graph n) : List Bool :=  (List.ofFn fun i : Fin n => List.ofFn fun j : Fin n => G.adj i j).flattendef encodePair {n : ℕ} (G H : Graph n) : List Bool :=  encodeNat n ++ matrixBits G ++ matrixBits Hinductive Move where  | left | stay | right  deriving DecidableEqdef Move.displacement : Move → ℤ  | .left => -1  | .stay => 0  | .right => 1structure TM where  tapes : ℕ  states : ℕ  symbols : ℕ  tapes_pos : 0 < tapes  states_ge : 3 ≤ states  symbols_ge : 3 ≤ symbols  transition : Fin states → (Fin tapes → Fin symbols) →    Fin states × (Fin tapes → Fin symbols) × (Fin tapes → Move)namespace TMstructure Config (M : TM) where  state : Fin M.states  head : Fin M.tapes → ℤ  tape : Fin M.tapes → ℤ → Fin M.symbolsdef symbol (M : TM) (b : Bool) : Fin M.symbols :=  if b then ⟨2, by have := M.symbols_ge; omega⟩ else    ⟨1, by have := M.symbols_ge; omega⟩def initial (M : TM) (w : List Bool) : M.Config where  state := ⟨0, by have := M.states_ge; omega⟩  head := fun _ => 0  tape := fun a z => if a.val = 0 ∧ 0 ≤ z ∧ z.toNat < w.length then    M.symbol (w.getD z.toNat false) else ⟨0, by have := M.symbols_ge; omega⟩def halted (M : TM) (C : M.Config) : Prop := C.state.val = 1 ∨ C.state.val = 2instance (M : TM) (C : M.Config) : Decidable (M.halted C) := inferInstanceAs  (Decidable (C.state.val = 1 ∨ C.state.val = 2))def step (M : TM) (C : M.Config) : M.Config :=  if M.halted C then C else  let tr := M.transition C.state (fun a => C.tape a (C.head a))  { state := tr.1    head := fun a => C.head a + (tr.2.2 a).displacement    tape := fun a z => if z = C.head a then tr.2.1 a else C.tape a z }def run (M : TM) (w : List Bool) (t : ℕ) : M.Config := M.step^[t] (M.initial w)def halts (M : TM) (w : List Bool) : Prop := ∃ t, M.halted (M.run w t)noncomputable def time (M : TM) (w : List Bool) : ℕ := by  classical  exact if h : M.halts w then Nat.find h else 0def accepts (M : TM) (w : List Bool) : Prop :=  ∃ t, (M.run w t).state.val = 1def outputs (M : TM) (input output : List Bool) : Prop :=  M.halts input ∧ ∀ i : Fin output.length,    (M.run input (M.time input)).tape ⟨0, M.tapes_pos⟩ (i.val : ℤ) =      M.symbol (output.get i)end TMabbrev Word (w : ℕ) := Fin (2 ^ w)def wordBits {w : ℕ} (a : Word w) : List Bool :=  List.ofFn (fun i : Fin w => a.val.testBit i.val)def registerBits {R w : ℕ} (a : Fin R → Word w) : List Bool :=  (List.ofFn fun i : Fin R => wordBits (a i)).flattenstructure WordOperation (R : ℕ) where  apply : (w : ℕ) → (Fin R → Word w) → (Fin R → Word w)  implementation : TM  exponent : ℕ  constant : ℕ  computes : ∀ w a, implementation.outputs    (encodeNat w ++ registerBits a) (registerBits (apply w a))  cost : ∀ w (a : Fin R → Word w), implementation.time (encodeNat w ++ registerBits a) ≤    constant * (w + 1) ^ exponentinductive RAMInstruction (R P : ℕ) where  | halt (answer : Bool)  | jump (pc : Fin P)  | compute (op : WordOperation R) (next : RAMInstruction R P)  | load (dest addr : Fin R) (next : RAMInstruction R P)  | store (addr src : Fin R) (next : RAMInstruction R P)  | branchZero (reg : Fin R) (yes no : RAMInstruction R P)structure WordWidth where  bits : ℕ → ℕ  positive : ∀ n, 0 < bits n  bound : ℕ  logarithmic : ∀ n, bits n ≤ bound * (Nat.log2 (n + 2) + 1)structure RAM where  registers : ℕ  locations : ℕ  locations_pos : 0 < locations  width : WordWidth  program : Fin locations → RAMInstruction registers locations  input_fits : ∀ n (G H : Graph n),    (encodePair G H).length ≤ width.bits n * 2 ^ width.bits nnamespace RAMstructure Config (A : RAM) (n : ℕ) where  pc : Fin A.locations  registers : Fin A.registers → Word (A.width.bits n)  memory : Word (A.width.bits n) → Word (A.width.bits n)  answer : Option Booldef packedWord (w : ℕ) (input : List Bool) (a : Word w) : Word w :=  ⟨(∑ i ∈ Finset.range w, if input.getD (w * a.val + i) false then 2 ^ i else 0)      % (2 ^ w), Nat.mod_lt _ (by positivity)⟩def initial (A : RAM) (n : ℕ) (w : List Bool) : A.Config n where  pc := ⟨0, A.locations_pos⟩  registers := fun _ => ⟨0, by positivity⟩  memory := packedWord (A.width.bits n) w  answer := nonedef execute (A : RAM) {n : ℕ} :    RAMInstruction A.registers A.locations → A.Config n → A.Config n  | .halt b, C => { C with answer := some b }  | .jump p, C => { C with pc := p }  | .compute op next, C => A.execute next { C with      registers := op.apply (A.width.bits n) C.registers }  | .load dest addr next, C => A.execute next { C with      registers := Function.update C.registers dest (C.memory (C.registers addr)) }  | .store addr src next, C => A.execute next { C with      memory := Function.update C.memory (C.registers addr) (C.registers src) }  | .branchZero reg yes no, C =>      if (C.registers reg).val = 0 then A.execute yes C else A.execute no Cdef step (A : RAM) {n : ℕ} (C : A.Config n) : A.Config n :=  match C.answer with  | some _ => C  | none => A.execute (A.program C.pc) Cdef run (A : RAM) (n : ℕ) (w : List Bool) (t : ℕ) : A.Config n :=  A.step^[t] (A.initial n w)def halts (A : RAM) (n : ℕ) (w : List Bool) : Prop :=  ∃ t, (A.run n w t).answer.isSome = truenoncomputable def time (A : RAM) (n : ℕ) (w : List Bool) : ℕ := by  classical  exact if h : A.halts n w then Nat.find h else 0def accepts (A : RAM) (n : ℕ) (w : List Bool) : Prop :=  ∃ t, (A.run n w t).answer = some trueend RAMinductive Model where  | turing (machine : TM)  | ram (machine : RAM)def Model.halts (A : Model) (n : ℕ) (w : List Bool) : Prop := match A with  | .turing M => M.halts w  | .ram M => M.halts n wdef Model.accepts (A : Model) (n : ℕ) (w : List Bool) : Prop := match A with  | .turing M => M.accepts w  | .ram M => M.accepts n wnoncomputable def Model.time (A : Model) (n : ℕ) (w : List Bool) : ℕ := match A with  | .turing M => M.time w  | .ram M => M.time n winductive InputClass where  | unrestricted | diameterTwo  deriving DecidableEqdef InputClass.contains (p : InputClass) {n : ℕ} (G H : Graph n) : Prop := match p with  | .unrestricted => True  | .diameterTwo => G.diameterTwo ∧ H.diameterTwodef Decides (A : Model) (c : Convention) (k : ℕ) (p : InputClass) : Prop :=  ∀ n (G H : Graph n), p.contains G H →    A.halts n (encodePair G H) ∧      (A.accepts n (encodePair G H) ↔ Equivalent c k G H)noncomputable def worstTime (A : Model) (p : InputClass) (n : ℕ) : ℕ := by  classical  exact Finset.univ.sup (fun pair : Graph n × Graph n =>    if p.contains pair.1 pair.2 then A.time n (encodePair pair.1 pair.2) else 0)def MainClaim : Prop :=  ∃ c : ℝ, 0 < c ∧ ∃ k₀ : ℕ, ∀ k : ℕ, k₀ ≤ k →    ∀ convention : Convention, ∀ p : InputClass, ∀ A : Model,      Decides A convention k p → ∃ n₀ : ℕ, ∀ n : ℕ, n₀ ≤ n →        (n : ℝ) ^ (c * (k : ℝ)) ≤ (worstTime A p n : ℝ)end WLTimetheorem WLTime.unconditional_time_lower_bound : WLTime.MainClaim := 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.

Parity lifts and bounded-treewidth witnesses for Weisfeiler–Leman equivalence

Theorem (Parity reduction)

Fix k\ge4, either refinement convention, and the corresponding t in (eq:bag-parameter). From every choice system on t domains one can explicitly construct two uncolored simple graphs X_0,X_\star of equal order such that

X_0\equiv_k X_\star \quad\Longleftrightarrow\quad \text{the choice system has no successful choice}.

The assertion holds both with common color names on separate graphs and with pure-tuple comparison in their disjoint union.

If every domain and label set has size at most A, where A\ge1, their common order is at most A M_t, where

M_t=t\,2^{\,t-2+3\binom{t-1}{2}}+12\binom t3.

When a successful choice exists, the two graphs already have different histograms after at most two joint rounds, or one separate round.

Section 1, Introduction Read it in the paper TeX source, line 76
Show the TeX
Fix \(k\ge4\), either refinement convention, and the corresponding
\(t\) in \eqref{eq:bag-parameter}. From every choice system on \(t\)
domains one can explicitly construct two uncolored simple graphs
\(X_0,X_\star\) of equal order such that
\[
 X_0\equiv_k X_\star
 \quad\Longleftrightarrow\quad
 \text{the choice system has no successful choice}.
\]
The assertion holds both with common color names on separate graphs
and with pure-tuple comparison in their disjoint union.

If every domain and label set has size at most \(A\), where \(A\ge1\),
their common order is at most \(A M_t\), where
\[
 M_t=t\,2^{\,t-2+3\binom{t-1}{2}}+12\binom t3.
\]
When a successful choice exists, the two graphs already have different
histograms after at most two joint rounds, or one separate round.
Theorem (Fixed-dimension ETH consequence)

Assume positive exponential-rate ETH. There are constants c>0 and K\ge4 such that, for every fixed integer k\ge K, no deterministic algorithm decides k-WL equivalence of two n-vertex graphs in time O(n^{ck}). This holds for both joint and separate refinement. The inputs are explicit adjacency matrices of uncolored simple undirected graphs, and time is measured on a multitape Turing machine.

Section 1, Introduction Read it in the paper TeX source, line 197
Show the TeX
Assume positive exponential-rate ETH. There are constants \(c>0\)
and \(K\ge4\) such that, for every fixed integer \(k\ge K\), no
deterministic algorithm decides \(k\)-WL equivalence of two
\(n\)-vertex graphs in time \(O(n^{ck})\).
This holds for both joint and separate refinement. The inputs are
explicit adjacency matrices of uncolored simple undirected graphs,
and time is measured on a multitape Turing machine.

The complexity of identifying a graph by Weisfeiler–Leman refinement

Theorem

The following problem is EXPTIME-complete under deterministic polynomial-time many-one reductions: given a nonempty finite simple undirected uncolored graph G by its adjacency matrix and a positive integer k in binary, decide whether \WLdim(G)\le k. Here dimension one uses ordinary neighbor color refinement, and dimensions at least two use the joint-update rule described above.

Section 1, Introduction Read it in the paper TeX source, line 36
Show the TeX
The following problem is EXPTIME-complete under deterministic
polynomial-time many-one reductions: given a nonempty finite simple
undirected uncolored graph $G$ by its adjacency matrix and a positive
integer $k$ in binary, decide whether $\WLdim(G)\le k$.
Here dimension one uses ordinary neighbor color refinement, and
dimensions at least two use the joint-update rule described above.

Unconditional time lower bounds for Weisfeiler–Leman equivalence

Theorem

There are absolute constants c>0 and k_0 with the following property. For every fixed integer k\ge k_0 and every correct deterministic decider A for k-Weisfeiler–Leman equivalence, there is an integer n_0=n_0(A,k) such that

T_A(n)\ge n^{ck} \qquad\text{for every integer }n\ge n_0.

This holds for multitape Turing machines and for sequential random-access machines with O(\log(n+2))-bit addresses and words, boundedly many memory accesses per instruction, and fixed instructions computable in time polynomial in the word length. The same bound holds if A is required to decide correctly only on pairs of simple connected uncolored graphs of diameter at most two, with T_A(n) maximized over that class. It applies both to the joint replacement-multiset convention for k-dimensional refinement and to the separate-multiset convention defined below.

Section 1, Introduction Read it in the paper TeX source, line 82
Show the TeX
There are absolute constants $c>0$ and $k_0$ with the following
property. For every fixed integer $k\ge k_0$ and every correct
deterministic decider $A$ for $k$-Weisfeiler--Leman equivalence, there
is an integer $n_0=n_0(A,k)$ such that
\[
                     T_A(n)\ge n^{ck}
             \qquad\text{for every integer }n\ge n_0.
\]
This holds for multitape Turing machines and for sequential
random-access machines with $O(\log(n+2))$-bit addresses and words,
boundedly many memory accesses per instruction, and fixed instructions
computable in time polynomial in the word length.
The same bound holds if $A$ is required to decide correctly only on
pairs of simple connected uncolored graphs of diameter at most two,
with $T_A(n)$ maximized over that class. It applies both to the
joint replacement-multiset convention for $k$-dimensional refinement
and to the separate-multiset convention defined below.

Variable-dimension Weisfeiler–Leman equivalence on general and subcubic graphs

Theorem

Both \WL and \SubWL are \EXPTIME-complete under deterministic polynomial-time many-one reductions.

Section 1, Introduction Read it in the paper TeX source, line 28
Show the TeX
Both \(\WL\) and \(\SubWL\) are \(\EXPTIME\)-complete under deterministic
polynomial-time many-one reductions.
Configuration and file facts
OpenAI’s Lean catalogue
OpenAI’s Lean catalogue does not list this as a main result; the release’s Lean notes for family 133 link it.
Solution module
OAI.Computability.WeisfeilerLeman.Main
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
52 local vocabulary declarations
Lines of Lean, without comments or blank lines
196
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

Unconditional sequential time lower bounds for Weisfeiler–Leman equivalence

Family 133: The computational complexity of Weisfeiler–Leman refinement

Candidate pairing

Theorem

There are absolute constants c>0 and k_0 with the following property. For every fixed integer k\ge k_0 and every correct deterministic decider A for k-Weisfeiler–Leman equivalence, there is an integer n_0=n_0(A,k) such that

T_A(n)\ge n^{ck} \qquad\text{for every integer }n\ge n_0.

This holds for multitape Turing machines and for sequential random-access machines with O(\log(n+2))-bit addresses and words, boundedly many memory accesses per instruction, and fixed instructions computable in time polynomial in the word length. The same bound holds if A is required to decide correctly only on pairs of simple connected uncolored graphs of diameter at most two, with T_A(n) maximized over that class. It applies both to the joint replacement-multiset convention for k-dimensional refinement and to the separate-multiset convention defined below.

Unconditional time lower bounds for Weisfeiler–Leman equivalence

TeX label thm:main, build/main.tex, line 82

The Lean statement says

def MainClaim : Prop :=  ∃ c : ℝ, 0 < c ∧ ∃ k₀ : ℕ, ∀ k : ℕ, k₀ ≤ k →    ∀ convention : Convention, ∀ p : InputClass, ∀ A : Model,      Decides A convention k p → ∃ n₀ : ℕ, ∀ n : ℕ, n₀ ≤ n →        (n : ℝ) ^ (c * (k : ℝ)) ≤ (worstTime A p n : ℝ)theorem WLTime.unconditional_time_lower_bound : WLTime.MainClaim := 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.

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

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

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