Efficiently trading off exploration and exploitation is one of the key challenges in online Reinforcement Learning (RL). Most works achieve this by carefully estimating the model uncertainty and following the so-called optimistic model. Inspired by practical ensemble methods, in this work we propose a simple and novel batch ensemble scheme that provably achieves near-optimal regret for stochastic Multi-Armed Bandits (MAB). Crucially, our algorithm has just a single parameter, namely the number of batches, and its value does not depend on distributional properties such as the scale and variance of the losses. We complement our theoretical results by demonstrating the effectiveness of our algorithm on synthetic benchmarks.
论文检索
输入标题、作者或关键词，从 559 篇学术成果中精准定位
559篇论文匹配“Online Learning and Bandits”
第 8 / 28 页
We study online fair division when there are a finite number of item types and the player values for the items are drawn randomly from distributions with unknown means. In this setting, a sequence of indivisible items arrives according to a random online process, and each item must be allocated to a single player. The goal is to maximize expected social welfare while maintaining that the allocation satisfies proportionality in expectation. When player values are normalized, we show that it is possible to with high probability guarantee proportionality constraint satisfaction and achieve O(√T) regret. To achieve this result, we present an upper confidence bound (UCB) algorithm that uses two rounds of linear optimization. This algorithm highlights fundamental aspects of proportionality constraints that allow for a UCB algorithm despite the presence of many (potentially tight) constraints. This result improves upon the previous best regret rate of O(T^(2/3)).
Coalition formation concerns autonomous agents that strategically interact to form self-organized coalitions. When agents lack initial sufficient information to evaluate their preferences before interacting with others, they learn them online through repeated feedback while iteratively forming coalitions. In this work, we introduce online learning in coalition formation from a non-cooperative perspective, studying the impact of collective data utilization where selfish agents aim to accelerate their learning by leveraging a shared data platform. Thus, the efficiency and dynamics of the learning process are affected by each agent's local feedbacks, motivating us to explore the tension between semi-bandit and bandit feedback, which differ in the granularity of utility information observed by each agent. Under our non-cooperative viewpoint, we evaluate the system by means of Nash stability, where no agent can improve her utility by unilaterally deviating. Our main result is a sample-efficient algorithm for selfish agents that aims to minimize their Nash regret under both semi-bandit and bandit feedback, implying approximately Nash stable outcomes. Under both feedback settings, our algorithm enjoys Nash regret and sample complexity bounds that are optimal up to logarithmic factors.
Large Language Models (LLMs) demonstrate impressive capabilities in the domain of program synthesis. This level of performance is not, however, universal across all tasks, all LLMs and all prompting styles. There are many areas where one LLM dominates, one prompting style dominates, or where calling a symbolic solver is a better choice than an LLM. A key challenge for the user then, is to identify not only when an LLM is the right choice of solver, and the appropriate LLM to call for a given synthesis task, but also the right way to call it. A non-expert user who makes the wrong choice, incurs a cost both in terms of results (number of tasks solved, and the time it takes to solve them) and financial cost, if using a closed-source language model via a commercial API. We frame this choice as an online learning problem. We use a multi-armed bandit algorithm to select which symbolic solver, or LLM and prompt combination to deploy in order to maximize a given reward function (which may prioritize solving time, number of synthesis tasks solved, or financial cost of solving). We implement an instance of this approach, called \name, and evaluate it on synthesis queries from the literature in ranking function synthesis, from the syntax-guided synthesis competition, and fresh, unseen queries generated from SMT problems. Cyanea solves 37.2 % more queries than the best single solver and achieves results within 4 % of the virtual best solver.
Large Language Models (LLMs) have revolutionized natural language processing, but their varying capabilities and costs pose challenges in practical applications. LLM routing addresses this by dynamically selecting the most suitable LLM for each query/task. Previous approaches treat this as a supervised learning problem, assuming complete knowledge of optimal query-LLM pairings. However, real-world scenarios lack such comprehensive mappings and face evolving user queries. We thus propose to study LLM routing as a contextual bandit problem, enabling adaptive decision-making using bandit feedback without requiring exhaustive inference across all LLMs for all queries (in contrast to supervised routing). To address this problem, we develop a shared embedding space for queries and LLMs, where query and LLM embeddings are aligned to reflect their affinity. This space is initially learned from offline human preference data and refined through online bandit feedback. We instantiate this idea through Preference-prior Informed Linucb fOr adaptive rouTing (PILOT), a novel extension of LinUCB. To handle diverse user budgets for model routing, we introduce an online cost policy modeled as a multi-choice knapsack problem, ensuring resource-efficient routing.
Theory · Reinforcement Learning and Planning
Many high-dimensional online decision-making problems can be modeled as stochastic sparse linear bandits. Most existing algorithms are designed to achieve optimal worst-case regret in either the data-rich regime, where polynomial dependence on the ambient dimension is unavoidable, or the data-poor regime, where dimension-independence is possible at the cost of worse dependence on the number of rounds. In contrast, the Bayesian approach of Information Directed Sampling (IDS) achieves the best of both worlds: a Bayesian regret bound that has the optimal rate in both regimes simultaneously. In this work, we explore the use of Sparse Optimistic Information Directed Sampling (SOIDS) to achieve the best of both worlds in the worst-case setting, without Bayesian assumptions. Through a novel analysis that enables the use of a time-dependent learning rate, we show that OIDS can be tuned without prior knowledge to optimally balance information and regret. Our results extend the theoretical guarantees of IDS, providing the first algorithm that simultaneously achieves optimal worst-case regret in both the data-rich and data-poor regimes. We empirically demonstrate the good performance of SOIDS.
Theory · Reinforcement Learning and Planning
We present regret minimization algorithms for the contextual multi-armed bandit (CMAB) problem over $K$ actions in the presence of delayed feedback, a scenario where loss observations arrive with delays chosen by an adversary. As a preliminary result, assuming direct access to a finite policy class $\Pi$ we establish an optimal expected regret bound of $ O (\sqrt{KT \log |\Pi|} + \sqrt{D \log |\Pi|)} $ where $D$ is the sum of delays. For our main contribution, we study the general function approximation setting over a (possibly infinite) contextual loss function class $ \mathcal{F} $ with access to an online least-square regression oracle $\mathcal{O}$ over $\mathcal{F}$. In this setting, we achieve an expected regret bound of $O(\sqrt{KTR_T(\mathcal{O})} + \sqrt{ d_{\max} D \beta})$ assuming FIFO order, where $d_{\max}$ is the maximal delay, $R_T(\mathcal{O})$ is an upper bound on the oracle's regret and $\beta$ is a stability parameter associated with the oracle. We complement this general result by presenting a novel stability analysis of a Hedge-based version of Vovk's aggregating forecaster as an oracle implementation for least-square regression over a finite function class $\mathcal{F}$ and show that its stability parameter $\beta$ is bounded by $\log |\mathcal{F}|$, resulting in an expected regret bound of $O(\sqrt{KT \log |\mathcal{F}|} + \sqrt{d_{\max} D \log |\mathcal{F}|})$ which is a $\sqrt{d_{\max}}$ factor away from the lower bound of $\Omega(\sqrt{KT \log |\mathcal{F}|} + \sqrt{D \log |\mathcal{F}|})$ that we also present.
Deep Learning · Everything Else
Early-Exit Deep Neural Networks enable adaptive inference by allowing prediction at intermediary layers, significantly reducing computational costs and latency. Most of the early exit strategies greedily exit a sample at an intermediary layer if the confidence in class prediction exceeds a predefined threshold that is set using a static validation set. This is problematic as the model might be overconfident in a wrong class. Also, they are not robust to distribution shifts encountered in deployment, which can undermine model trustworthiness and accuracy. To address these challenges, we propose UAT that adapts the threshold for exit decisions using a Multi-Armed Bandit framework, enabling online, unsupervised adjustment of exit decisions. UAT makes decisions based on a new reward function that assesses predictive certainty and its reliability to balance computational efficiency and prediction quality while penalizing unnecessary late exits. We provide guarantees on risk achieved by UAT and validate its performance on diverse tasks spanning vision-language understanding, text generation, and classification. Our framework demonstrates consistent improvements in speedup $(1.70-2.10\times)$ with a minimal performance drop $(<2)$\% as compared to full model performance.
Theory · Reinforcement Learning and Planning
Efficient online decision-making in contextual bandits is challenging, as methods without informative priors often suffer from computational or statistical inefficiencies. In this work, we leverage pre-trained diffusion models as expressive priors to capture complex action dependencies and develop a practical algorithm that efficiently approximates posteriors under such priors, enabling both fast updates and sampling. Empirical results demonstrate the effectiveness and versatility of our approach across diverse contextual bandit settings.
Probabilistic Methods
We introduce scalable algorithms for online learning of neural network parameters and Bayesian sequential decision making. Unlike classical Bayesian neural networks, which induce predictive uncertainty through a posterior over model parameters, our methods adopt a predictive-first perspective based on martingale posteriors. In particular, we work directly with the one-step-ahead posterior predictive, which we parameterize with a neural network and update sequentially with incoming observations. This decouples Bayesian decision-making from parameter-space inference: we sample from the posterior predictive for decision making, and update the parameters of the posterior predictive via fast, frequentist Kalman-filter-like recursions. Our algorithms operate in a fully online, replay-free setting, providing principled uncertainty quantification without costly posterior sampling. Empirically, they achieve competitive performance–speed trade-offs in non-stationary contextual bandits and Bayesian optimization, offering 10–100 times faster inference than classical Thompson sampling while maintaining comparable or superior decision performance.
Theory · Reinforcement Learning and Planning
We study the generalized linear bandit (GLB) problem, a contextual multi-armed bandit framework that extends the classical linear model by incorporating a non-linear link function, thereby modeling a broad class of reward distributions such as Bernoulli and Poisson. While GLBs are widely applicable to real-world scenarios, their non-linear nature introduces significant challenges in achieving both computational and statistical efficiency. Existing methods typically trade off between two objectives, either incurring high per-round costs for optimal regret guarantees or compromising statistical efficiency to enable constant-time updates. In this paper, we propose a jointly efficient algorithm that attains a nearly optimal regret bound with $\mathcal{O}(1)$ time and space complexities per round. The core of our method is a tight confidence set for the online mirror descent (OMD) estimator, which is derived through a novel analysis that leverages the notion of mix loss from online prediction. The analysis shows that our OMD estimator, even with its one-pass updates, achieves statistical efficiency comparable to maximum likelihood estimation, thereby leading to a jointly efficient optimistic method.
Theory · Reinforcement Learning and Planning
We study online learning in finite-horizon episodic Markov decision processes (MDPs) under the challenging \textit{aggregate bandit feedback} model, where the learner observes only the cumulative loss incurred in each episode, rather than individual losses at each state-action pair. While prior work in this setting has focused exclusively on worst-case analysis, we initiate the study of \textit{best-of-both-worlds} (BOBW) algorithms that achieve low regret in both stochastic and adversarial environments. We propose the first BOBW algorithms for episodic tabular MDPs with aggregate bandit feedback. In the case of known transitions, our algorithms achieve $O(\log T)$ regret in stochastic settings and ${O}(\sqrt{T})$ regret in adversarial ones. Importantly, we also establish matching lower bounds, showing the optimality of our algorithms in this setting. We further extend our approach to unknown-transition settings by incorporating confidence-based techniques. Our results rely on a combination of FTRL over occupancy measures, self-bounding techniques, and new loss estimators inspired by recent advances in online shortest path problems. Along the way, we also provide the first individual-gap-dependent lower bounds and demonstrate near-optimal BOBW algorithms for shortest path problems with bandit feedback.
Theory · Reinforcement Learning and Planning
The problem of contextual dueling bandits is central to reinforcement learning with human feedback (RLHF), a widely used approach in AI alignment for incorporating human preferences into learning systems. Despite its importance, existing methods are constrained either by strong preference modeling assumptions or by applicability only to finite action spaces. Moreover, prior algorithms typically rely on online optimization oracles, which are computationally infeasible for complex function classes, limiting their practical effectiveness. In this work, we present the first fundamental theoretical study of general contextual dueling bandits over continuous action spaces. Our key contribution is a novel algorithm based on a regularized min-max optimization framework that achieves a regret bound of $\tilde{O}(\sqrt{dT})$—the first such guarantee for this general setting. By leveraging offline oracles instead of online ones, our method further improves computational efficiency. Empirical evaluations validate our theoretical findings, with our approach significantly outperforming existing baselines in terms of regret.
Reinforcement Learning · Multi-agent
We explore the use of expert-guided bandit learning, which we refer to as online mixture-of-experts (OMoE). In this setting, given a context, a candidate committee of experts must determine how to aggregate their outputs to achieve optimal results in terms of aggregate accuracy. We propose two algorithms to address this problem. The first algorithm combines aggregate voting with UCB-driven successive elimination, efficiently pruning suboptimal exploration actions. The second algorithm employs an online weighted-majority-voting mechanism, leveraging the respective voting power of each expert proportional to their predictive power. We derive theoretical guarantees for the regret properties in the bandit setting under ideal circumstances, and empirical results are provided accordingly. As a modern study on applications, these methods are applied to the online fine-tuning of a set of expert large language models (LLMs), where after each response, the generative LLM dynamically reweighs its set of experts and/or selects the optimal committee of experts to generate the most accurate response. Our results introduce new methodologies and no-regret guarantees for combining multiple experts to improve on the performance of the an aggregate model overall.
Theory · Reinforcement Learning and Planning
We revisit the contextual cascading bandit, where a learning agent recommends an ordered list (\emph{cascade}) of items, and a user scans the list sequentially, stopping at the first attractive item. Although cascading bandits underpin various applications including recommender systems and search engines, the role of the cascade length $K$ in shaping regret has remained unclear. Contrary to prior results that regret grows with $K$, we prove that regret actually \emph{decreases} once $K$ is large enough. Leveraging this insight, we design a new upper-confidence-bound algorithm built on online mirror descent that attains the sharpest known regret upper bound, $\tilde{\mathcal{O}}\bigl(\min \lbrace K\bar{p}^{K-1}, 1 \rbrace d \sqrt{T}\bigr)$ for contextual cascading bandits. To complement this new regret upper bound, we provide a nearly matching lower bound of $\Omega \bigl(\min \lbrace K\underline{p}^{K-1}, 1 \rbrace d \sqrt{T}\bigr)$, where $0 \leq \underline{p} \leq \bar{p} < 1$. Together, these results fully characterize how regret truly scales with $K$, thereby closing the theoretical gap for contextual cascading bandits. Finally, comprehensive experiments validate our theoretical results and show the effectiveness of our proposed method.
Theory · Learning Theory
We study the Pandora’s Box problem in an online learning setting with semi-bandit feedback. In each round, the learner sequentially pays to open up to $n$ boxes with unknown reward distributions, observes rewards upon opening, and decides when to stop. The utility of the learner is the maximum observed reward minus the cumulative cost of opened boxes, and the goal is to minimize regret defined as the gap between the cumulative expected utility and that of the optimal policy. We propose a new algorithm that achieves $\widetilde{O}(\sqrt{nT})$ regret after $T$ rounds, which improves the $\widetilde{O}(n\sqrt{T})$ bound of Agarwal et al. [2024] and matches the known lower bound up to logarithmic factors. To better capture real-life applications, we then extend our results to a natural but challenging contextual linear setting, where each box's expected reward is linear in some known but time-varying $d$-dimensional context and the noise distribution is fixed over time. We design an algorithm that learns both the linear function and the noise distributions, achieving $\widetilde{O}(nd\sqrt{T})$ regret. Finally, we show that our techniques also apply to the online Prophet Inequality problem, where the learner must decide immediately whether or not to accept a revealed reward. In both non-contextual and contextual settings, our approach achieves similar improvements and regret bounds.
Theory · Learning Theory
We study the problem of multiclass learning with bandit feedback in both the i.i.d. batch and adversarial online models. In the *uniform* learning framework, it is well known that no hypothesis class $\mathcal{H}$ is learnable in either model when the effective number of labels is unbounded. In contrast, within the *universal* learning framework, recent works by (Hanneke et al., 2025b) and (Hanneke et al., 2025a) have established surprising exact equivalences between learnability under bandit feedback and full supervision in both the i.i.d. batch and adversarial online models, respectively. This raises a natural question: What happens in the *non-uniform* learning framework, which lies between the uniform and universal learning frameworks? Our contributions are twofold: (1) We provide a combinatorial characterization of learnable hypothesis classes in both models, in the realizable and agnostic settings, within the non-uniform learning framework. Notably, this includes elementary and natural hypothesis classes, such as a countably infinite collection of constant functions over some domain that is learnable in both models. (2) We construct a hypothesis class that is non-uniformly learnable under full supervision in the adversarial online model (and thus also in the i.i.d. batch model), but not non-uniformly learnable under bandit feedback in the i.i.d. batch model (and thus also not in the adversarial online model). This serves as our main novel technical contribution that reveals a fundamental distinction between the non-uniform and universal learning frameworks.
Theory · Reinforcement Learning and Planning
Many online learning applications adopt the stochastic bandit problem with a linear reward model, where the unknown parameter exhibits a succinct structure. We study minimax regret lower bounds which allow to know whether more efficient algorithms can be proposed. We introduce a general definition of succinctness and propose a novel framework for constructing minimax regret lower bounds based on an information-regret trade-off. When applied to entry-sparse vectors, our framework sharpens a recent lower bound by (Hao et al, NeurIPS 2020). We further apply our framework to derive novel results. To the best of our knowledge, we provide the first lower bounds for the group-sparse and low-rank matrix settings.
Theory · Learning Theory
We study online convex optimization on $\ell_p$-balls in $\mathbb{R}^d$ for $p > 2$. While always sub-linear, the optimal regret exhibits a shift between the high-dimensional setting ($d > T$), when the dimension $d$ is greater than the time horizon $T$ and the low-dimensional setting ($d \leq T$). We show that Follow-the-Regularised-Leader (FTRL) with time-varying regularisation which is adaptive to the dimension regime is anytime optimal for all dimension regimes. Motivated by this, we ask whether it is possible to obtain anytime optimality of FTRL with fixed non-adaptive regularisation. Our main result establishes that for separable regularisers, adaptivity in the regulariser is necessary, and that any fixed regulariser will be sub-optimal in one of the two dimension regimes. Finally, we provide lower bounds which rule out sub-linear regret bounds for the linear bandit problem in sufficiently high-dimension for all $\ell_p$-balls with $p \geq 1$.
Theory · Learning Theory
The multi-armed bandit (MAB) is a fundamental online decision-making framework that has been extensively studied over the past two decades. To mitigate the high cost and slow convergence of purely online learning, modern MAB approaches have explored _hybrid_ paradigms that leverage offline data to warm-start online learning. However, existing approaches face a significant limitation by assuming that the offline and online data are homogeneous—they share the same feedback structure and are drawn from the same underlying distribution. This assumption is often violated in practice, where offline data often originate from diverse sources and evolving environments, resulting in feedback heterogeneity and distributional shifts. In this work, we tackle the challenge of learning across this offline-online gap by developing a general hybrid bandit framework that incorporates heterogeneous offline data to improve online performance. We study two hybrid settings: (1) using reward-based offline data to accelerate online learning in preference-based bandits (i.e., dueling bandits), and (2) using preference-based offline data to improve online standard MAB algorithms. For both settings, we design novel algorithms and derive tight regret bounds that match or improve upon existing benchmarks despite heterogeneity. Empirical evaluations on both synthetic and real-world datasets show that our proposed methods significantly outperform baseline algorithms.