AFTD Auto-Formalizing Theoretical Domains

c1_k_submod_kernel_proj

theoremverified

The first coordinate projection function (fun x _ => x) : ℝ → ℝ → ℝ satisfies is_c1_k_submod_kernel.

Statement

theorem c1_k_submod_kernel_proj : is_c1_k_submod_kernel (fun x _ => x)
Source
arXiv:2610.26677, Lemma 2.1
Verified
22 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 3
  • is_c1_k_submod_kernelA bivariate function H : ℝ → ℝ → ℝ satisfies is_c1_k_submod_kernel if the map (fun p : ℝ × ℝ => H p.1 p.2) is continuously differentiable…
  • k_submod_kernel_projThe coordinate projection function (fun x _ => x) : ℝ → ℝ → ℝ satisfies is_k_submod_kernel.
  • is_k_submod_kernelA two-variable kernel H : ℝ → ℝ → ℝ satisfies the hidden-label conditions if H(0,0) = 0, H(1,0) > 0, H(x,y) ≤ H(1,0) on [0,1]², H is…

Read back from the Lean

The projection function $H : ℝ → ℝ → ℝ$ defined by $H(x, y) = x$ (i.e., fun x _ => x) satisfies is_c1_k_submod_kernel, which means: 1. The uncurried function $(x, y) \mapsto x$ from $ℝ \times ℝ$ to $ℝ$ is continuously differentiable of class $C^1$ (ContDiff ℝ 1). 2. $H(0, 0) = 0$. 3. $0 < H(1, 0)$. 4. For all $x, y \in [0, 1]$, $H(x, y) \le H(1, 0)$. 5. For all $x_1, x_2, y \in ℝ$ with $0 \le x_1 \le x_2 \le 1$ and $0 \le y \le 1$, $H(x_1, y) \le H(x_2, y)$ (non-decreasing in the first argument). 6. For all $x, y_1, y_2 \in ℝ$ with $0 \le x \le 1$ and $0 \le y_1 \le y_2 \le 1$, $H(x, y_1) \le H(x, y_2)$ (non-decreasing in the second argument). 7. For all $x_1, x_2, y_1, y_2 \in ℝ$ with $0 \le x_1 \le x_2 \le 1$ and $0 \le y_1 \le y_2 \le 1$, the derivative of $t \mapsto H(t, y_2)$ evaluated at $x_2$ is less than or equal to the derivative of $t \mapsto H(t, y_1)$ evaluated at $x_1$. 8. For all $x_1, x_2, y_1, y_2 \in ℝ$ with $0 \le x_1 \le x_2 \le 1$ and $0 \le y_1 \le y_2 \le 1$, the derivative of $t \mapsto H(x_2, t)$ evaluated at $y_2$ is less than or equal to the derivative of $t \mapsto H(x_1, t)$ evaluated at $y_1$.

Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.

The proof, explained

The theorem establishes that the first coordinate projection (x, y) ↦ x is a C¹ submodular kernel. By definition, this requires showing that the bivariate projection is of class C¹ and that it satisfies the submodular kernel inequalities. The proof constructs a direct pair consisting of `contDiff_fst` (which establishes smoothness of the first projection) and `k_submod_kernel_proj` (which verifies the kernel conditions).

  1. The first coordinate projection (x, y) ↦ x is C¹-smooth and satisfies the conditions of a submodular kernel.

    By definition, `is_c1_k_submod_kernel H` is a conjunction of two conditions: the bivariate map (x, y) ↦ H x y must be of class C¹ on ℝ × ℝ, and H must satisfy the submodular kernel properties `is_k_submod_kernel H`. For H(x, y) = x, the bivariate map is the first coordinate projection Prod.fst, which is smooth of all orders (and hence C¹) by `contDiff_fst`. The submodularity conditions for this projection are provided by `k_submod_kernel_proj`. The proof simply pairs these two witnesses together.

    line 2

    uses contDiff_fst, k_submod_kernel_proj

    theorem c1_k_submod_kernel_proj : is_c1_k_submod_kernel (fun x _ => x) := ⟨contDiff_fst, k_submod_kernel_proj⟩
contDiff_fst
The first coordinate projection on a product of normed spaces is of class Cⁿ for any smoothness degree n.
k_submod_kernel_proj
The first coordinate projection function (x, y) ↦ x satisfies the submodular kernel conditions (is_k_submod_kernel).

Written by a model (gemini-3.8-flash) from the Lean source alone (no goal states were available). Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.

Lean source view module on GitHub

/-- The coordinate projection kernel satisfies is_c1_k_submod_kernel (sanity witness). -/theorem c1_k_submod_kernel_proj : is_c1_k_submod_kernel (fun x _ => x) := ⟨contDiff_fst, k_submod_kernel_proj⟩
See it in the graphDiscuss this resultChallenge it