arXiv subjects

Ilgam Latypov

Publications and source records attributed to Ilgam Latypov.

6 recordsLinked to original sources

arXiv2026

Managing Self-Learning Experts under Per-Round Budget Constraints

This paper addresses the problem of sequential decision-making under learning budget constraints. Such settings naturally arise in applications like managing a portfolio of bandit or reinforcement learning (RL) algorithms. We propose a novel UCB-type algorithm, M-LCB, designed to manage a pool of $K$ self-learning experts in a stochastic environment while accounting for a limited per-round learning budget $M$. At each round, M-LCB selects one expert to make a decision and at most $M \le K$ experts to learn. For selection, M-LCB uses confidence bounds constructed from limited prior knowledge about the experts (i.e., mild assumptions) and their observed training losses. We derive anytime regret bounds for M-LCB that scale with the individual regrets of the experts. In particular, if each expert has regret $\tilde O(T^α)$ by round $T$, then M-LCB guarantees an overall regret of $\tilde O\left(\sqrt{KT/M} + (K/M)^{1-α}T^α\right)$ relative to the best expert in hindsight. Finally, we demonstrate the applicability of M-LCB using self-learning experts instantiated as (i) parametric models and (ii) bandit algorithms.

cs.LG↗

arXiv2026

Localize, Restart, Accelerate: Stochastic Optimization under Generalized Smoothness

We study stochastic convex optimization under asymmetric $(L_0,L_1)$-generalized smoothness, a model motivated by machine-learning objectives whose local curvature may grow with the gradient norm. We assume an unbiased first-order oracle with additive norm-sub-Gaussian noise. Acceleration is difficult in this setting because momentum may enter regions of much larger curvature, while stochastic gradients cannot reliably certify an unrestricted trajectory. We propose \textsf{ARC-SG}, a two-phase accelerated method: Phase~I reduces excessively large gradients using a generalized-smoothness-aware stochastic step, then Phase~II solves strongly convex proximal subproblems by a restarted accelerated solver confined to certified smoothness balls. Exact proximal points do not increase the gradient norm, allowing these certificates to propagate through the outer loop. The contribution is a query-by-query certified-localization construction with explicit generalized-smoothness factors and a strongly convex restart extension. \textsf{ARC-SG} achieves, with high probability, an accelerated optimization contribution and smooth-subclass-optimal statistical dependence on accuracy, up to logarithmic and generalized-smoothness factors. Its convex accuracy exponents agree with a contemporaneous public stochastic-acceleration result under a broader smoothness and affine-variance model; our distinction is the certified geometry, explicit parameter accounting, and strongly convex guarantee. The results recover classical accelerated stochastic rates when $L_1=0$. Experiments on objectives with unbounded gradients illustrate the two-phase mechanism and its finite-budget advantage.

math.OC↗

arXiv2026

Functional multi-armed bandit and the best function identification problems

Bandit optimization usually refers to the class of online optimization problems with limited feedback, namely, a decision maker uses only the objective value at the current point to make a new decision and does not have access to the gradient of the objective function. While this name accurately captures the limitation in feedback, it is somehow misleading since it does not have any connection with the multi-armed bandits (MAB) problem class. We propose two new classes of problems: the functional multi-armed bandit problem (FMAB) and the best function identification problem. They are modifications of a multi-armed bandit problem and the best arm identification problem, respectively, where each arm represents an unknown black-box function. These problem classes are a surprisingly good fit for modeling real-world problems such as competitive LLM training. To solve the problems from these classes, we propose a new reduction scheme to construct UCB-type algorithms, namely, the F-LCB algorithm, based on algorithms for nonlinear optimization with known convergence rates. We provide the regret upper bounds for this reduction scheme based on the base algorithms' convergence rates. We add numerical experiments that demonstrate the performance of the proposed scheme.

cs.LG↗

arXiv2024

$γ$-Competitiveness: An Approach to Multi-Objective Optimization with High Computation Costs in Lipschitz Functions

In practical engineering and optimization, solving multi-objective optimization (MOO) problems typically involves scalarization methods that convert a multi-objective problem into a single-objective one. While effective, these methods often incur significant computational costs due to iterative calculations and are further complicated by the need for hyperparameter tuning. In this paper, we introduce an extension of the concept of competitive solutions and propose the Scalarization With Competitiveness Method (SWCM) for multi-criteria problems. This method is highly interpretable and eliminates the need for hyperparameter tuning. Additionally, we offer a solution for cases where the objective functions are Lipschitz continuous and can only be computed once, termed Competitiveness Approximation on Lipschitz Functions (CAoLF). This approach is particularly useful when computational resources are limited or re-computation is not feasible. Through computational experiments on the minimum-cost concurrent flow problem, we demonstrate the efficiency and scalability of the proposed method, underscoring its potential for addressing computational challenges in MOO across various applications.

math.OC↗

arXiv2024

EEvA: Fast Expert-Based Algorithms for Buffer Page Replacement

Optimal page replacement is an important problem in efficient buffer management. The range of replacement strategies known in the literature varies from simple but efficient FIFO-based algorithms to more accurate but potentially costly methods tailored to specific data access patterns. The principal issue in adopting a pattern-specific replacement logic in a DB buffer manager is to guarantee non-degradation in general high-load regimes. In this paper, we propose a new family of page replacement algorithms for DB buffer manager which demonstrate a superior performance wrt competitors on custom data access patterns and imply a low computational overhead on TPC-C. We provide theoretical foundations and an extensive experimental study on the proposed algorithms which covers synthetic benchmarks and an implementation in an open-source DB kernel evaluated on TPC-C.

cs.DB↗

arXiv2024

Fast UCB-type algorithms for stochastic bandits with heavy and super heavy symmetric noise

In this study, we propose a new method for constructing UCB-type algorithms for stochastic multi-armed bandits based on general convex optimization methods with an inexact oracle. We derive the regret bounds corresponding to the convergence rates of the optimization methods. We propose a new algorithm Clipped-SGD-UCB and show, both theoretically and empirically, that in the case of symmetric noise in the reward, we can achieve an $O(\log T\sqrt{KT\log T})$ regret bound instead of $O\left (T^{\frac{1}{1+α}} K^{\fracα{1+α}} \right)$ for the case when the reward distribution satisfies $\mathbb{E}_{X \in D}[|X|^{1+α}] \leq σ^{1+α}$ ($α\in (0, 1])$, i.e. perform better than it is assumed by the general lower bound for bandits with heavy-tails. Moreover, the same bound holds even when the reward distribution does not have the expectation, that is, when $α<0$.

cs.LG↗