Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

§4, proof of Theorem 7 — the increment of the Bregman divergences is at most ½‖x*−y‖₂² tr(G_{t+1}^{1/2}−G_t^{1/2})

Proved
AdaGrad.Full.display_bregman_increment

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

adagradmatrix-analysisp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1

Let g1,g2,⋯∈Rdg_1,g_2,\dots\in\mathbb R^dg1​,g2​,⋯∈Rd, Gt=∑τ≤tgτgτ⊤G_t=\sum_{\tau\le t}g_\tau g_\tau^\topGt​=∑τ≤t​gτ​gτ⊤​, δ∈R\delta\in\mathbb Rδ∈R, Ht=δI+Gt1/2H_t=\delta I+G_t^{1/2}Ht​=δI+Gt1/2​ and Bψt(x,y)=12⟨x−y,Ht(x−y)⟩B_{\psi_t}(x,y)=\tfrac12\langle x-y,H_t(x-y)\rangleBψt​​(x,y)=21​⟨x−y,Ht​(x−y)⟩. For every t≥0t\ge0t≥0 and all x∗,y∈Rdx^*,y\in\mathbb R^dx∗,y∈Rd,

Bψt+1(x∗,y)−Bψt(x∗,y)=12⟨x∗−y,(Gt+11/2−Gt1/2)(x∗−y)⟩≤12∥x∗−y∥22tr⁡(Gt+11/2−Gt1/2).B_{\psi_{t+1}}(x^*,y)-B_{\psi_t}(x^*,y)=\tfrac12\big\langle x^*-y,(G_{t+1}^{1/2}-G_t^{1/2})(x^*-y)\big\rangle\le\tfrac12\|x^*-y\|_2^2\operatorname{tr}\big(G_{t+1}^{1/2}-G_t^{1/2}\big).Bψt+1​​(x∗,y)−Bψt​​(x∗,y)=21​⟨x∗−y,(Gt+11/2​−Gt1/2​)(x∗−y)⟩≤21​∥x∗−y∥22​tr(Gt+11/2​−Gt1/2​).

Applied with y=xt+1y=x_{t+1}y=xt+1​ it bounds the middle sum of Proposition 3 for full-matrix ADAGRAD.

Formalization Note The page passes through 12∥x∗−y∥22λmax⁡(Gt+11/2−Gt1/2)\tfrac12\|x^*-y\|_2^2\lambda_{\max}(G_{t+1}^{1/2}-G_t^{1/2})21​∥x∗−y∥22​λmax​(Gt+11/2​−Gt1/2​); that intermediate quantity is not stated, only the equality and the outer inequality.

Preamble
import Mathlib
import Definitions.Def_AdaGrad_Full_Algorithm2
open scoped MatrixOrder InnerProductSpace
Formal statement
namespace AdaGrad.Full

/-- §4, proof of Theorem 7, first display (p. 2134): for ADAGRAD's proximal matrices
`H_t = δI + G_t^{1/2}`, `B_{ψ_{t+1}}(x*, y) − B_{ψ_t}(x*, y) = ½⟨x* − y, (G_{t+1}^{1/2} − G_t^{1/2})(x* − y)⟩`
and this is at most `½‖x* − y‖₂² tr(G_{t+1}^{1/2} − G_t^{1/2})`. -/
theorem display_bregman_increment {d : ℕ} (δ : ℝ) (g : ℕ → EuclideanSpace ℝ (Fin d)) (t : ℕ)
    (xstar y : EuclideanSpace ℝ (Fin d)) :
    bregman (H δ g (t + 1)) xstar y - bregman (H δ g t) xstar y
        = 1 / 2 * mInner (S g (t + 1) - S g t) (xstar - y) (xstar - y) ∧
      1 / 2 * mInner (S g (t + 1) - S g t) (xstar - y) (xstar - y)
        ≤ 1 / 2 * ‖xstar - y‖ ^ 2 * (S g (t + 1) - S g t).trace := by sorry

end AdaGrad.Full
Source
Duchi, Hazan, Singer, Adaptive Subgradient Methods for Online Learning and Stochastic Optimization, JMLR 12 (2011), p. 2134, §4, proof of Theorem 7, first display
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 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