When Lower-Order Terms Dominate: Adaptive Expert Algorithms for Heavy-Tailed Losses

• URL: http://arxiv.org/abs/2506.01722v1
• Date: Mon, 02 Jun 2025 14:29:05 GMT
• Title: When Lower-Order Terms Dominate: Adaptive Expert Algorithms for Heavy-Tailed Losses
• Authors: Antoine Moulin, Emmanuel Esposito, Dirk van der Hoeven,
• Abstract summary: We develop adaptive algorithms that do not require prior knowledge about the range or the second moment of the losses.<n>Existing adaptive algorithms have what is typically considered a lower-order term in their regret guarantees.
• Score: 12.39860047886679
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We consider the problem setting of prediction with expert advice with possibly heavy-tailed losses, i.e.\ the only assumption on the losses is an upper bound on their second moments, denoted by $\theta$. We develop adaptive algorithms that do not require any prior knowledge about the range or the second moment of the losses. Existing adaptive algorithms have what is typically considered a lower-order term in their regret guarantees. We show that this lower-order term, which is often the maximum of the losses, can actually dominate the regret bound in our setting. Specifically, we show that even with small constant $\theta$, this lower-order term can scale as $\sqrt{KT}$, where $K$ is the number of experts and $T$ is the time horizon. We propose adaptive algorithms with improved regret bounds that avoid the dependence on such a lower-order term and guarantee $\mathcal{O}(\sqrt{\theta T\log(K)})$ regret in the worst case, and $\mathcal{O}(\theta \log(KT)/\Delta_{\min})$ regret when the losses are sampled i.i.d.\ from some fixed distribution, where $\Delta_{\min}$ is the difference between the mean losses of the second best expert and the best expert. Additionally, when the loss function is the squared loss, our algorithm also guarantees improved regret bounds over prior results.

Related papers

• Exploiting Curvature in Online Convex Optimization with Delayed Feedback [6.390468088226495]
  We study the online convex optimization problem with curved losses and delayed feedback.<n>We propose a variant of follow-the-regularized-leader that obtains regret of order $minsigma_maxln T, sqrtd_mathrmtot$.<n>We then consider exp-concave losses and extend the Online Newton Step algorithm to handle delays with an adaptive learning rate tuning.
  arXiv  Detail & Related papers  (2025-06-09T09:49:54Z)
• Near-optimal Regret Using Policy Optimization in Online MDPs with Aggregate Bandit Feedback [49.84060509296641]
  We study online finite-horizon Markov Decision Processes with adversarially changing loss and aggregate bandit feedback (a.k.a full-bandit)<n>Under this type of feedback, the agent observes only the total loss incurred over the entire trajectory, rather than the individual losses at each intermediate step within the trajectory.<n>We introduce the first Policy Optimization algorithms for this setting.
  arXiv  Detail & Related papers  (2025-02-06T12:03:24Z)
• An Optimistic Algorithm for Online Convex Optimization with Adversarial Constraints [55.2480439325792]
  We study Online Convex Optimization (OCO) with adversarial constraints.<n>We focus on a setting where the algorithm has access to predictions of the loss and constraint functions.<n>Our results show that we can improve the current best bounds of $ O(sqrtT) $ regret and $ tildeO(sqrtT) $ cumulative constraint violations.
  arXiv  Detail & Related papers  (2024-12-11T03:06:42Z)
• Optimal Multiclass U-Calibration Error and Beyond [31.959887895880765]
  We consider the problem of online multiclass bounds U-calibration, where a forecaster aims to make sequential distributional predictions over $K$ classes with low U-calibration error. We show that the optimal U-calibration error is $Theta(sqrtKT)$.
  arXiv  Detail & Related papers  (2024-05-28T20:33:18Z)
• Improved Regret for Bandit Convex Optimization with Delayed Feedback [50.46856739179311]
  bandit convex optimization (BCO) with delayed feedback, where only the loss value of the action is revealed under a delay. We develop a novel algorithm, and prove that it enjoys a regret bound of $O(sqrtnT3/4+sqrtdT)$ in general. We show that the proposed algorithm can improve the regret bound to $O((nT)2/3log/3T+dlog T)$ for strongly convex functions.
  arXiv  Detail & Related papers  (2024-02-14T13:08:26Z)
• Non-stationary Online Convex Optimization with Arbitrary Delays [50.46856739179311]
  This paper investigates the delayed online convex optimization (OCO) in non-stationary environments. We first propose a simple algorithm, namely DOGD, which performs a gradient descent step for each delayed gradient according to their arrival order. We develop an improved algorithm, which reduces those dynamic regret bounds achieved by DOGD to $O(sqrtbardT(P_T+1))$.
  arXiv  Detail & Related papers  (2023-05-20T07:54:07Z)
• First- and Second-Order Bounds for Adversarial Linear Contextual Bandits [22.367921675238318]
  We consider the adversarial linear contextual bandit setting, which allows for the loss functions associated with each of $K$ arms to change over time without restriction. Since $V_T$ or $L_T*$ may be significantly smaller than $T$, these improve over the worst-case regret whenever the environment is relatively benign.
  arXiv  Detail & Related papers  (2023-05-01T14:00:15Z)
• Private Online Prediction from Experts: Separations and Faster Rates [74.52487417350221]
  Online prediction from experts is a fundamental problem in machine learning and several works have studied this problem under privacy constraints. We propose and analyze new algorithms for this problem that improve over the regret bounds of the best existing algorithms for non-adaptive adversaries.
  arXiv  Detail & Related papers  (2022-10-24T18:40:19Z)
• Constant regret for sequence prediction with limited advice [0.0]
  We provide a strategy combining only p = 2 experts per round for prediction and observing m $ge$ 2 experts' losses. If the learner is constrained to observe only one expert feedback per round, the worst-case regret is the "slow rate" $Omega$($sqrt$ KT)
  arXiv  Detail & Related papers  (2022-10-05T13:32:49Z)
• Private Stochastic Convex Optimization: Optimal Rates in $\ell_1$ Geometry [69.24618367447101]
  Up to logarithmic factors the optimal excess population loss of any $(varepsilon,delta)$-differently private is $sqrtlog(d)/n + sqrtd/varepsilon n.$ We show that when the loss functions satisfy additional smoothness assumptions, the excess loss is upper bounded (up to logarithmic factors) by $sqrtlog(d)/n + (log(d)/varepsilon n)2/3.
  arXiv  Detail & Related papers  (2021-03-02T06:53:44Z)
• Adapting to Delays and Data in Adversarial Multi-Armed Bandits [7.310043452300736]
  We analyze variants of the Exp3 algorithm that tune their step-size using only information available at the time of the decisions. We obtain regret guarantees that adapt to the observed (rather than the worst-case) sequences of delays and/or losses.
  arXiv  Detail & Related papers  (2020-10-12T20:53:52Z)
• Taking a hint: How to leverage loss predictors in contextual bandits? [63.546913998407405]
  We study learning in contextual bandits with the help of loss predictors. We show that the optimal regret is $mathcalO(minsqrtT, sqrtmathcalETfrac13)$ when $mathcalE$ is known.
  arXiv  Detail & Related papers  (2020-03-04T07:36:38Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.