subadditive_f_weight_hasDerivAt_facts
If f is C⁴ on (0, ∞), then at every x > 0, f, f', g = x² f'' and g' have derivatives f', f'', g', g'' at x.
Statement
theorem subadditive_f_weight_hasDerivAt_facts (f : ℝ → ℝ) (hsmooth : ContDiffOn ℝ 4 f (Set.Ioi 0)) :
∀ x > (0:ℝ), HasDerivAt f (deriv f x) x ∧ HasDerivAt (deriv f) (deriv (deriv f) x) x ∧
HasDerivAt (subadditive_f_weight f) (deriv (subadditive_f_weight f) x) x ∧
HasDerivAt (deriv (subadditive_f_weight f)) (deriv (deriv (subadditive_f_weight f)) x) x- Source
- arXiv:2610.21231 (On the subadditivity of mutual f-information), Appendix B, proof of Lemma 4
- Verified
- 24 Oct 2026
- Axioms
Classical.choiceQuot.soundpropext- Built on 1
- subadditive_f_weightThe weight g(x) = x² f''(x) attached to a generator f in Theorem 1 of arXiv:2610.21231 (f'' is the iterated derivative
deriv (deriv f)).
- subadditive_f_weightThe weight g(x) = x² f''(x) attached to a generator f in Theorem 1 of arXiv:2610.21231 (f'' is the iterated derivative
- Used by 1
- mutual_f_theta_concaveOn_of_weight_boundLemma 4 of arXiv:2610.21231: under the assumptions of Theorem 1 (g = x² f'' with g'' < 0 and g + 2(g' - η)²/(-g'') ≤ 2η + (x-1)ℓ), for…
The proof, explained
To show that f, f', g(x) = x² f''(x), and g' are all differentiable at x > 0 with derivatives given by the deriv operator, the proof first establishes their higher-order smoothness on (0, ∞). Differentiating the C⁴ hypothesis on f successively shows that f' is C³ and f'' is C² on (0, ∞). The function g is C² as the product of x ↦ x² and f'', which further implies that g' is C¹. Since each of these four functions is at least C¹ on the open neighbourhood (0, ∞) of x, each is differentiable at x, establishing the four derivative relations.
Fix x > 0 and note that the open interval (0, ∞) is a neighbourhood of x.
We introduce the point x and the assumption hx : x > 0. Since x > 0, the open ray (0, ∞) is a neighbourhood of x by `Ioi_mem_nhds`.
uses
Ioi_mem_nhdsintro x hx have hn : Set.Ioi (0:ℝ) ∈ nhds x := Ioi_mem_nhds hx
The first derivative f' is C³ and the second derivative f'' is C² on (0, ∞).
Because (0, ∞) is an open set (`isOpen_Ioi`) and f is C⁴ on (0, ∞), differentiation lowers the smoothness order by 1, yielding that f' is C³ since 3 + 1 ≤ 4. Repeating this argument for f' on the open set (0, ∞) gives that f'' is C² since 2 + 1 ≤ 3. The index inequalities are checked by `norm_num`.
uses
isOpen_Ioihave h1 : ContDiffOn ℝ 3 (deriv f) (Set.Ioi 0) := hsmooth.deriv_of_isOpen isOpen_Ioi (by norm_num) have h2 : ContDiffOn ℝ 2 (deriv (deriv f)) (Set.Ioi 0) := h1.deriv_of_isOpen isOpen_Ioi (by norm_num)
The weighted function g(x) = x² f''(x) is C² on (0, ∞).
By definition, `subadditive_f_weight f` is the function x ↦ x² f''(x). The monomial x ↦ x² is smooth everywhere (derived from `contDiffOn_id`), and f'' is C² on (0, ∞). Since the product of two C² functions is C², g is C² on (0, ∞).
uses
contDiffOn_idhave hg : ContDiffOn ℝ 2 (subadditive_f_weight f) (Set.Ioi 0) := by have : subadditive_f_weight f = fun x => x ^ 2 * deriv (deriv f) x := rfl rw [this] exact (contDiffOn_id.pow 2).mul h2
The derivative g' is C¹ on (0, ∞).
Because (0, ∞) is open (`isOpen_Ioi`) and g is C² on (0, ∞), differentiating g yields that g' is C¹ on (0, ∞), with the condition 1 + 1 ≤ 2 verified by `norm_num`.
uses
isOpen_Ioihave hg1 : ContDiffOn ℝ 1 (deriv (subadditive_f_weight f)) (Set.Ioi 0) := hg.deriv_of_isOpen isOpen_Ioi (by norm_num)
At x, each of f, f', g, and g' is differentiable with derivative given by the deriv operator.
The goal splits into the four derivative claims. Each of the four functions (f ∈ C⁴, f' ∈ C³, g ∈ C², and g' ∈ C¹) has smoothness order at least 1, which guarantees differentiability on the set (0, ∞). Because (0, ∞) is a neighbourhood of x, differentiability on the set implies differentiability at the point x, which in turn implies that the function has derivative equal to its deriv value at x.
refine ⟨?_, ?_, ?_, ?_⟩ · exact ((hsmooth.differentiableOn (by norm_num)).differentiableAt hn).hasDerivAt · exact ((h1.differentiableOn (by norm_num)).differentiableAt hn).hasDerivAt · exact ((hg.differentiableOn (by norm_num)).differentiableAt hn).hasDerivAt · exact ((hg1.differentiableOn (by norm_num)).differentiableAt hn).hasDerivAt
Ioi_mem_nhds- In a linear order with the order topology, the open ray (a, ∞) is a neighbourhood of any point b > a.
isOpen_Ioi- The open ray (a, ∞) is an open set.
contDiffOn_id- The identity function is C^n on any set, for any smoothness order n.
- In Lean, `deriv f x` is a total function returning the derivative when it exists (and 0 otherwise). The relation `HasDerivAt f (deriv f x) x` formally asserts that `f` is differentiable at `x` and that its derivative equals `deriv f x`.
- The tactic `norm_num` is used repeatedly to discharge routine arithmetic on smoothness indices (such as 3 + 1 ≤ 4, 2 + 1 ≤ 3, 1 + 1 ≤ 2, and n ≠ 0).
Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.
Lean source view module on GitHub
Choose a tactic — a marked line — to see the goals Lean had before and after it, as in Lean’s infoview. Step with ← → or j k.
theorem subadditive_f_weight_hasDerivAt_facts (f : ℝ → ℝ) (hsmooth : ContDiffOn ℝ 4 f (Set.Ioi 0)) : ∀ x > (0:ℝ), HasDerivAt f (deriv f x) x ∧ HasDerivAt (deriv f) (deriv (deriv f) x) x ∧ HasDerivAt (subadditive_f_weight f) (deriv (subadditive_f_weight f) x) x ∧ HasDerivAt (deriv (subadditive_f_weight f)) (deriv (deriv (subadditive_f_weight f)) x) x := by intro x hx have hn : Set.Ioi (0:ℝ) ∈ nhds x := Ioi_mem_nhds hx have h1 : ContDiffOn ℝ 3 (deriv f) (Set.Ioi 0) := hsmooth.deriv_of_isOpen isOpen_Ioi (by norm_num) have h2 : ContDiffOn ℝ 2 (deriv (deriv f)) (Set.Ioi 0) := h1.deriv_of_isOpen isOpen_Ioi (by norm_num) have hg : ContDiffOn ℝ 2 (subadditive_f_weight f) (Set.Ioi 0) := by have : subadditive_f_weight f = fun x => x ^ 2 * deriv (deriv f) x := rfl rw [this] exact (contDiffOn_id.pow 2).mul h2 have hg1 : ContDiffOn ℝ 1 (deriv (subadditive_f_weight f)) (Set.Ioi 0) := hg.deriv_of_isOpen isOpen_Ioi (by norm_num) refine ⟨?_, ?_, ?_, ?_⟩ · exact ((hsmooth.differentiableOn (by norm_num)).differentiableAt hn).hasDerivAt · exact ((h1.differentiableOn (by norm_num)).differentiableAt hn).hasDerivAt · exact ((hg.differentiableOn (by norm_num)).differentiableAt hn).hasDerivAt · exact ((hg1.differentiableOn (by norm_num)).differentiableAt hn).hasDerivAt