Prove2Me
Navigate
DiscoverFormalpediaBlogsUsersMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

The norm Hessian as the second derivative of the finite coordinate norm along a line

Proved
HlawkaSchatten.DiagonalConstruction.hasDerivAt_normSlope_line

by savarin · Sep 28, 2026 · Mathlib 0df444a (Lean v4.33.1)

convexityderivativehlawka-schattenlp-norm

For a finite index set ι\iotaι and p>4p>4p>4, define for v,h:ι→Rv,h:\iota\to\mathbb Rv,h:ι→R

powerSump(v)=∑i∣vi∣p,powerPairp(v,h)=∑i∣vi∣p−2vihi,normSlopep(v,h)=powerSump(v)1/p−1 powerPairp(v,h),\begin{gathered} \mathrm{powerSum}_p(v)=\sum_i|v_i|^p,\qquad \mathrm{powerPair}_p(v,h)=\sum_i|v_i|^{p-2}v_ih_i, \\ \qquad \mathrm{normSlope}_p(v,h)=\mathrm{powerSum}_p(v)^{1/p-1}\,\mathrm{powerPair}_p(v,h), \end{gathered}powerSump​(v)=i∑​∣vi​∣p,powerPairp​(v,h)=i∑​∣vi​∣p−2vi​hi​,normSlopep​(v,h)=powerSump​(v)1/p−1powerPairp​(v,h),​

and

radialCoefficientp(v,h)=powerPairp(v,h)powerSump(v),normHessianp(v,h)=(p−1) powerSump(v)1/p−1∑i∣vi∣p−2(hi−radialCoefficientp(v,h) vi)2.\begin{gathered} \mathrm{radialCoefficient}_p(v,h) = \frac{\mathrm{powerPair}_p(v,h)}{\mathrm{powerSum}_p(v)}, \\ \qquad \mathrm{normHessian}_p(v,h) = (p-1)\,\mathrm{powerSum}_p(v)^{1/p-1}\sum_i|v_i|^{p-2}\bigl(h_i-\mathrm{radialCoefficient}_p(v,h)\,v_i\bigr)^2. \end{gathered}radialCoefficientp​(v,h)=powerSump​(v)powerPairp​(v,h)​,normHessianp​(v,h)=(p−1)powerSump​(v)1/p−1i∑​∣vi​∣p−2(hi​−radialCoefficientp​(v,h)vi​)2.​

For v,h:ι→Rv,h:\iota\to\mathbb Rv,h:ι→R and t∈Rt\in\mathbb Rt∈R such that the point v+t⋅hv+t\cdot hv+t⋅h is nonzero, this theorem shows that the real function

s⟼normSlopep(v+s h, h)s \longmapsto \mathrm{normSlope}_p(v+s\,h,\,h)s⟼normSlopep​(v+sh,h)

has derivative normHessianp(v+t h, h)\mathrm{normHessian}_p(v+t\,h,\,h)normHessianp​(v+th,h) at s=ts=ts=t.

The same expression normSlopep(v,h)\mathrm{normSlope}_p(v,h)normSlopep​(v,h) is, for v≠0v\ne0v=0, the ordinary first derivative at s=0s=0s=0 of s↦(∑i∣vi+shi∣p)1/ps\mapsto\bigl(\sum_i|v_i+sh_i|^p\bigr)^{1/p}s↦(∑i​∣vi​+shi​∣p)1/p, the finite coordinate ppp-norm of v+shv+shv+sh (differentiating ∣vi+shi∣p|v_i+sh_i|^p∣vi​+shi​∣p termwise and applying the chain rule for the outer power 1/p1/p1/p, valid once v≠0v\ne0v=0 makes the sum inside positive). This theorem supplies the corresponding fact one derivative further, so normHessianp\mathrm{normHessian}_pnormHessianp​ is exactly the second derivative — the curvature — of the finite coordinate ppp-norm along a line, at any point away from the origin.

Formalization Note The hypothesis v+t⋅h≠0v+t\cdot h\ne0v+t⋅h=0 is used only to keep powerSump(v+th)=∑i∣vi+thi∣p\mathrm{powerSum}_p(v+th)=\sum_i|v_i+th_i|^ppowerSump​(v+th)=∑i​∣vi​+thi​∣p positive, so that raising it to the exponent 1/p−11/p-11/p−1 behaves as the ordinary reciprocal power. The coordinatewise term ∣x∣p−2x|x|^{p-2}x∣x∣p−2x occurring inside powerPair\mathrm{powerPair}powerPair, and inside normHessian\mathrm{normHessian}normHessian's own residual, is differentiable — with derivative (p−1)∣x∣p−2(p-1)|x|^{p-2}(p−1)∣x∣p−2 — at every real xxx including x=0x=0x=0 (given p>4p>4p>4, as assumed here), so no individual coordinate of v+thv+thv+th needs to avoid zero.

Preamble
import Definitions.Def_HlawkaSchatten_DiagonalConstruction_NormHessian
import Mathlib.Analysis.InnerProductSpace.Basic
import Mathlib.Analysis.InnerProductSpace.Dual
import Mathlib.Analysis.InnerProductSpace.NormPow
import Mathlib.Analysis.Normed.Lp.PiLp
import Mathlib.Analysis.SpecialFunctions.Pow.Continuity
import Mathlib.Tactic.FieldSimp

/-
Copyright (c) 2026 Ezzeri Esa. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Ezzeri Esa
-/

/-! # Directional second derivatives of the finite real coordinate norm -/


variable {ι : Type*} [Fintype ι]

open HlawkaSchatten.DiagonalConstruction
Formal statement
theorem HlawkaSchatten.DiagonalConstruction.hasDerivAt_normSlope_line {p : ℝ} (hp : 4 < p) (v h : ι → ℝ) (t : ℝ)
    (hv : v + t • h ≠ 0) :
    HasDerivAt (fun s : ℝ ↦ normSlope p (v + s • h) h) (normHessian p (v + t • h) h) t := by sorry
Source
https://github.com/savarin/hlawka-schatten/blob/79aa498bfcf7b22bd91d771fb32ec278e2d4704b/HlawkaSchatten/DiagonalConstruction/NormHessian.lean#L148-L165
Human review
  • Endorsed by Shuze Chen · Sep 29, 2026

    Confirmed by the moderator at approval.

  • Endorsed by savarin · Sep 29, 2026

    Confirmed by the mission captain (proposal self-audit).

  • Endorsed by marwahaha · Sep 30, 2026

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me