Median Clipping for Zeroth-order Non-Smooth Convex Optimization and Multi-Armed Bandit Problem with Heavy-tailed Symmetric Noise

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kornilov, Nikita, Dorn, Yuriy, Lobanov, Aleksandr, Kutuzov, Nikolay, Shibaev, Innokentiy, Gorbunov, Eduard, Nazin, Alexander, Gasnikov, Alexander
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911302665895936
author Kornilov, Nikita
Dorn, Yuriy
Lobanov, Aleksandr
Kutuzov, Nikolay
Shibaev, Innokentiy
Gorbunov, Eduard
Nazin, Alexander
Gasnikov, Alexander
author_facet Kornilov, Nikita
Dorn, Yuriy
Lobanov, Aleksandr
Kutuzov, Nikolay
Shibaev, Innokentiy
Gorbunov, Eduard
Nazin, Alexander
Gasnikov, Alexander
contents In this paper, we consider non-smooth convex optimization with a zeroth-order oracle corrupted by symmetric stochastic noise. Unlike the existing high-probability results requiring the noise to have bounded $κ$-th moment with $κ\in (1,2]$, our results allow even heavier noise with any $κ> 0$, e.g., the noise distribution can have unbounded expectation. Our convergence rates match the best-known ones for the case of the bounded variance, namely, to achieve function accuracy $\varepsilon$ our methods with Lipschitz oracle require $\tilde{O}(d^2\varepsilon^{-2})$ iterations for any $κ> 0$. We build the median gradient estimate with bounded second moment as the mini-batched median of the sampled gradient differences. We apply this technique to the stochastic multi-armed bandit problem with heavy-tailed distribution of rewards and achieve $\tilde{O}(\sqrt{dT})$ regret. We demonstrate the performance of our zeroth-order and MAB algorithms for various $κ\in (0,2]$ on synthetic and real-world data. Our methods do not lose to SOTA approaches and dramatically outperform them for $κ\leq 1$.
format Preprint
id arxiv_https___arxiv_org_abs_2402_02461
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Median Clipping for Zeroth-order Non-Smooth Convex Optimization and Multi-Armed Bandit Problem with Heavy-tailed Symmetric Noise
Kornilov, Nikita
Dorn, Yuriy
Lobanov, Aleksandr
Kutuzov, Nikolay
Shibaev, Innokentiy
Gorbunov, Eduard
Nazin, Alexander
Gasnikov, Alexander
Optimization and Control
In this paper, we consider non-smooth convex optimization with a zeroth-order oracle corrupted by symmetric stochastic noise. Unlike the existing high-probability results requiring the noise to have bounded $κ$-th moment with $κ\in (1,2]$, our results allow even heavier noise with any $κ> 0$, e.g., the noise distribution can have unbounded expectation. Our convergence rates match the best-known ones for the case of the bounded variance, namely, to achieve function accuracy $\varepsilon$ our methods with Lipschitz oracle require $\tilde{O}(d^2\varepsilon^{-2})$ iterations for any $κ> 0$. We build the median gradient estimate with bounded second moment as the mini-batched median of the sampled gradient differences. We apply this technique to the stochastic multi-armed bandit problem with heavy-tailed distribution of rewards and achieve $\tilde{O}(\sqrt{dT})$ regret. We demonstrate the performance of our zeroth-order and MAB algorithms for various $κ\in (0,2]$ on synthetic and real-world data. Our methods do not lose to SOTA approaches and dramatically outperform them for $κ\leq 1$.
title Median Clipping for Zeroth-order Non-Smooth Convex Optimization and Multi-Armed Bandit Problem with Heavy-tailed Symmetric Noise
topic Optimization and Control
url https://arxiv.org/abs/2402.02461