Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 5 — robust regression has a solution supported on III if some allowable perturbation of the features outside III makes them irrelevant

Proved
RobustRegLasso.Sparsity.theorem_5_sparse_solution

by mikedeng1 · Oct 7, 2026 · Mathlib 0df444a (Lean v4.33.1)

lassop2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1regressionrobust-optimizationsparsity

Let a1,…,am,b∈Rna_1,\dots,a_m, b \in \mathbb R^na1​,…,am​,b∈Rn, radii ci≥0c_i \ge 0ci​≥0, and U={(δ1,…,δm)∣∥δi∥2≤ci}\mathcal U = \{(\delta_1,\dots,\delta_m) \mid \|\delta_i\|_2 \le c_i\}U={(δ1​,…,δm​)∣∥δi​∥2​≤ci​}, the feature-wise uncoupled uncertainty set. Let I⊆{1,…,m}I \subseteq \{1,\dots,m\}I⊆{1,…,m}, and write UI={ΔAI∣ΔA∈U}\mathcal U^I = \{\Delta A^I \mid \Delta A \in \mathcal U\}UI={ΔAI∣ΔA∈U}, where ΔAI\Delta A^IΔAI agrees with ΔA\Delta AΔA on the features in III and is zero elsewhere.

Suppose there is a perturbation ΔA~Ic∈U\Delta\tilde A^{I^c} \in \mathcal UΔA~Ic∈U of the features in IcI^cIc (that is, zero on every feature in III) such that the robust regression problem

min⁡x∈Rm{max⁡ΔA~I∈UI∥b−(A+ΔA~Ic+ΔA~I)x∥2}\min_{x\in\mathbb R^m}\Big\{\max_{\Delta\tilde A^I\in\mathcal U^I}\|b-(A+\Delta\tilde A^{I^c}+\Delta\tilde A^I)x\|_2\Big\}x∈Rmmin​{ΔA~I∈UImax​∥b−(A+ΔA~Ic+ΔA~I)x∥2​}

has a solution supported on III. Then the robust regression problem

min⁡x∈Rm{max⁡ΔA∈U∥b−(A+ΔA)x∥2}\min_{x\in\mathbb R^m}\Big\{\max_{\Delta A\in\mathcal U}\|b-(A+\Delta A)x\|_2\Big\}x∈Rmmin​{ΔA∈Umax​∥b−(A+ΔA)x∥2​}

has a solution supported on III: some optimal solution x∗x^*x∗ has xj∗=0x^*_j = 0xj∗​=0 for all j∉Ij \notin Ij∈/I.

In words, a robust regression has an optimal solution supported on III if some allowable perturbation of the features outside III makes them irrelevant. This connects robustness to sparsity, and is the tool from which the paper derives its incoherence-type criterion (Theorem 6).

Formalization Note "Has a solution supported on III" is the existence of a minimizer of the robust objective (the EReal supremum over the set) with zero entries off III. The perturbation ΔA~Ic\Delta\tilde A^{I^c}ΔA~Ic is encoded as an element of U\mathcal UU that vanishes on III. The radii ci≥0c_i \ge 0ci​≥0 are the paper's standing convention. The paper writes I⊆{1,…,n}I \subseteq\{1,\dots,n\}I⊆{1,…,n} on p. 8; features are indexed by 1,…,m1,\dots,m1,…,m.

Preamble
import Mathlib
import Definitions.Def_RobustRegLasso_Sparsity_Basic
Formal statement
namespace RobustRegLasso.Sparsity

theorem theorem_5_sparse_solution {n m : ℕ} (a : Fin m → EuclideanSpace ℝ (Fin n))
    (b : EuclideanSpace ℝ (Fin n)) (c : Fin m → ℝ) (hc : ∀ i, 0 ≤ c i) (I : Finset (Fin m))
    (h : ∃ δTilde ∈ (RobustRegLasso.FeatureWise.uncertaintySet c : Set (Fin m → EuclideanSpace ℝ (Fin n))),
      (∀ i, i ∈ I → δTilde i = 0) ∧
      HasSolutionSupportedOn (a + δTilde) b (restrictedSet c I) I) :
    HasSolutionSupportedOn a b (RobustRegLasso.FeatureWise.uncertaintySet c) I := by sorry

end RobustRegLasso.Sparsity
Source
Xu, Caramanis, Mannor, Robust Regression and Lasso, arXiv:0811.1790v1, p. 8, Theorem 5
Human review
  • Endorsed by Shuze Chen · Oct 7, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 7, 2026

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

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 paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me