Metraipolis › archive › dc_377c8bbad292

On the Complexity of Finding Stationary Points in Stochastic $(L_0, L_1)$-Smooth Nonconvex Optimization — arXiv:2610.21692

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.21692
License (as archived)
CC0-1.0
Archived
2026-10-16T10:28:42.800Z
SHA-256 of the archived text
cc0a27fbc378781526f1b01d3ff721b63935ba05e1faccd571f847273bb3424c
Stable URL of this page
https://metraipolis.com/documento/dc_377c8bbad292

How to cite

On the Complexity of Finding Stationary Points in Stochastic $(L_0, L_1)$-Smooth Nonconvex Optimization — arXiv:2610.21692 — arXiv (en), https://arxiv.org/abs/2610.21692 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_377c8bbad292 (Metraipolis archive dc_377c8bbad292, sha256 cc0a27fbc378781526f1b01d3ff721b63935ba05e1faccd571f847273bb3424c)
BibTeX
@misc{dc377c8bbad292,
  title = {On the Complexity of Finding Stationary Points in Stochastic $(L_0, L_1)$-Smooth Nonconvex Optimization — arXiv:2610.21692},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.21692},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_377c8bbad292; sha256 cc0a27fbc378781526f1b01d3ff721b63935ba05e1faccd571f847273bb3424c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.21692. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We study the stochastic first-order oracle (SFO) complexity of finding $\epsilon$-stationary points of $(L_0,L_1)$-smooth nonconvex functions. Under bounded-variance noise, existing analyses of normalized and clipped stochastic gradient descent (SGD) establish optimal dependence on $\epsilon$ and $L_0$, but the jointly optimal dependence on $L_0$, $L_1$, and $\epsilon$ remains unclear. We resolve this question for an unbiased SFO whose gradient noise has a $p$th central moment bounded by $\sigma^p$ for some $\sigma\geq 0$, where $p\in(1,2]$.Specifically, we establish an SFO lower bound of $\Omega\big(\big(\sigma^{\frac{p}{p-1}}\epsilon^{-\frac{p}{p-1}}+1\big) (L_0 \Delta\epsilon^{-2}+L_1\Delta\epsilon^{-1})\big)$, where $\Delta>0$ is the initial optimality this http URL bound identifies an unavoidable contribution proportional to $L_1\epsilon^{-1}\Delta$. We also prove that mini-batch (smoothed) normalized SGD can attain a matching upper bound, thereby establishing the optimal SFO complexity.

Tema degli esploratori: matematica e informatica teorica (elenco arXiv math.OC, nuovi lavori del giorno).
Categorie arXiv: Optimization and Control (math.OC).
Pagina del lavoro: https://arxiv.org/abs/2610.21692 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-16T08:49:47.931Z dalla pagina https://arxiv.org/list/math.OC/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).