Adversarial attacks on stochastic bandits have traditionally relied on some unrealistic assumptions, such as per-round reward manipulation and unbounded perturbations, limiting their relevance to real-world systems. We propose a more practical threat model, Fake Data Injection, which reflects realistic adversarial constraints: the attacker can inject only a limited number of bounded fake feedback samples into the learner's history, simulating legitimate interactions. We design effective attack strategies under this model, explicitly addressing both magnitude constraints (on reward values) and temporal constraints (on when and how often data can be injected). Our theoretical analysis shows that these attacks can mislead a class of bandit algorithms into selecting a target arm in nearly all rounds while incurring only sublinear attack cost. Experiments on synthetic and real-world datasets validate the effectiveness of our strategies, revealing vulnerabilities in stochastic bandit algorithms under practical adversarial scenarios.
We study the problem of $K$-armed bandits with reward distributions belonging to a one-parameter exponential distribution family. In the literature, several criteria have been proposed to evaluate the performance of such algorithms, including Asymptotic Optimality, Minimax Optimality, Sub-UCB, and variance-adaptive regret bound. Thompson Sampling-based and Upper Confidence Bound-based algorithms have been employed to achieve some of these criteria. However, none of these algorithms simultaneously satisfy all the aforementioned criteria. In this paper, we design an algorithm, Exponential Kullback-LeiblerMaillard Sampling (abbrev. Exp-KL-MS), that achieves multiple optimality criteria simultaneously, including Asymptotic Optimality, minimax ratio with a $\sqrt{\ln (K)}$ factor, Sub-UCB, and a variance-adaptive regret bound. Our algorithm design follows the Minimum Empirical Divergence framework~\citep{honda2011asymptotically,maillard2011apprentissage}, with the exploration probability of arm $a$ proportional to $\text{exp}\left(-L(N_{t-1, a}) \text{KL}(\hat{\mu}_{t-1, a}, \max_{a'} \hat{\mu}_{t-1, a'})\right)$, where $L(\cdot)$ is an inverse temperature function, $N_{t-1, a}$ is the number of times arm $a$ has been pulled before time $t$, $\hat{\mu}_{t-1, a}$ is the empirical mean of arm $a$ before time $t$, and $\text{KL}(\cdot, \cdot)$ is the Kullback-Leibler divergence between two distributions in the one-parameter exponential distribution family. Our analysis allows different choices of inverse temperature function $L(k)$. We also provide numerical simulations demonstrating the effectiveness of our algorithms.
The increasing automation of navigation for unmanned aerial vehicles (UAVs) has exposed them to adversarial attacks that exploit vulnerabilities in reinforcement learning (RL) through sensor manipulation. Although existing robust RL methods aim to mitigate such threats, their effectiveness has limited generalization to out-of-distribution shifts from the optimal value distribution, as they are primarily designed to handle fixed perturbation. To address this limitation, this paper introduces an antifragile RL framework that enhances adaptability to broader distributional shifts by incorporating a switching mechanism based on discounted Thompson sampling (DTS). This mechanism dynamically selects among multiple robust policies to minimize adversarially induced state-action-value distribution shifts. The proposed approach first derives a diverse ensemble of action robust policies by accounting for a range of perturbations in the policy space. These policies are then modeled as a multiarmed bandit (MAB) problem, where DTS optimally selects policies in response to nonstationary Bernoulli rewards, effectively adapting to evolving adversarial strategies. Theoretical framework has also been provided where by optimizing the DTS to minimize the overall regrets due to distributional shift, results in effective adaptation against unseen adversarial attacks thus inducing antifragility. Extensive numerical simulations validate the effectiveness of the proposed framework in complex navigation environments with multiple dynamic three-dimensional obstacles and with stronger projected gradient descent (PGD) and spoofing attacks. Compared to conventional robust, non-adaptive RL methods, the antifragile approach achieves superior performance, demonstrating shorter navigation path lengths and a higher rate of conflict-free navigation trajectories compared to existing robust RL techniques
Bayesian optimization (BO) is a widely used iterative algorithm for optimizing black-box functions. Each iteration requires maximizing an acquisition function, such as the upper confidence bound (UCB) or a sample path from the Gaussian process (GP) posterior, as in Thompson sampling (TS). However, finding an exact solution to these maximization problems is often intractable and computationally expensive. Reflecting such realistic situations, in this paper, we delve into the effect of inexact maximizers of the acquisition functions. Defining a measure of inaccuracy in acquisition solutions, we establish cumulative regret bounds for both GP-UCB and GP-TS without requiring exact solutions of acquisition function maximization. Our results show that under appropriate conditions on accumulated inaccuracy, inexact BO algorithms can still achieve sublinear cumulative regret. Motivated by such findings, we provide both theoretical justification and numerical validation for random grid search as an effective and computationally efficient acquisition function solver.
Understanding which parts of the retrieved context contribute to a large language model's generated answer is essential for building interpretable and trustworthy generative QA systems. We propose a novel framework that formulates context attribution as a combinatorial multi-armed bandit (CMAB) problem. Each context segment is treated as a bandit arm, and we employ Combinatorial Thompson Sampling (CTS) to efficiently explore the exponentially large space of context subsets under a limited query budget. Our method defines a reward function based on normalized token likelihoods, capturing how well a subset of segments supports the original model response. Unlike traditional perturbation-based attribution methods such as SHAP, which sample subsets uniformly and incur high computational costs, our approach adaptively balances exploration and exploitation by leveraging posterior estimates of segment relevance. This leads to substantially improved query efficiency while maintaining high attribution fidelity. Extensive experiments on diverse datasets and LLMs demonstrate that our method achieves competitive attribution quality with fewer model queries.
In linear contextual bandits, the objective is to select actions that maximize cumulative rewards, modeled as a linear function with unknown parameters. Although Thompson Sampling performs well empirically, it does not achieve optimal regret bounds. This paper proposes a nearly minimax optimal Thompson Sampling for linear contextual bandits by developing a novel estimator with the adaptive augmentation and coupling of the hypothetical samples that are designed for efficient parameter learning. The proposed estimator accurately predicts rewards for all arms without relying on assumptions for the context distribution. Empirical results show robust performance and significant improvement over existing methods.
We consider a kernelized bandit problem with a compact arm set ${X} \subset \mathbb{R}^d $ and a fixed but unknown reward function $f^*$ with a finite norm in some Reproducing Kernel Hilbert Space (RKHS). We propose a class of computationally efficient kernelized bandit algorithms, which we call GP-Generic, based on a novel concept: exploration distributions. This class of algorithms includes Upper Confidence Bound-based approaches as a special case, but also allows for a variety of randomized algorithms. With careful choice of exploration distribution, our proposed generic algorithm realizes a wide range of concrete algorithms that achieve $\tilde{O}(\gamma_T\sqrt{T})$ regret bounds, where $\gamma_T$ characterizes the RKHS complexity. This matches known results for UCB- and Thompson Sampling-based algorithms; we also show that in practice, randomization can yield better practical results.
The challenge of identifying the best feasible arm within a fixed budget has attracted considerable interest in recent years. However, a notable gap remains in the literature: the exact exponential rate at which the error probability approaches zero has yet to be established, even in the relatively simple setting of $K$-armed bandits with Gaussian noise. In this paper, we address this gap by examining the problem within the context of linear bandits. We introduce a novel algorithm for best feasible arm identification that guarantees an exponential decay in the error probability. Remarkably, the decay rate -- characterized by the exponent -- matches the theoretical lower bound derived using information-theoretic principles. Our approach leverages a posterior sampling framework embedded within a game-based sampling rule involving a min-learner and a max-learner. This strategy shares its foundations with Thompson sampling, but is specifically tailored to optimize the identification process under fixed-budget constraints. Furthermore, we validate the effectiveness of our algorithm through comprehensive empirical evaluations across various problem instances with different levels of complexity. The results corroborate our theoretical findings and demonstrate that our method outperforms several benchmark algorithms in terms of both accuracy and efficiency.
Reinforcement learning from human feedback (RLHF) has achieved great empirical success in aligning large language models (LLMs) with human preference, and it is of great importance to study the statistical efficiency of RLHF algorithms from a theoretical perspective. In this work, we consider the online RLHF setting where the preference data is revealed during the learning process and study action value function approximation. We design a model-free posterior sampling algorithm for online RLHF inspired by Thompson sampling and provide its theoretical guarantee. Specifically, we adopt Bellman eluder (BE) dimension as the complexity measure of the function class and establish $O(\sqrt{T})$ regret bound for the proposed algorithm with other multiplicative factor depending on the horizon, BE dimension and the $log$-bracketing number of the function class. Further, in the analysis, we first establish the concentration-type inequality of the squared Bellman error bound based on the maximum likelihood estimator (MLE) generalization bound, which plays the crucial rules in obtaining the eluder-type regret bound and may be of independent interest.
The optimization of expensive black-box functions is ubiquitous in science and engineering. A common solution to this problem is Bayesian optimization (BO), which is generally comprised of two components: (i) a surrogate model and (ii) an acquisition function, which generally require expensive re-training and optimization steps at each iteration, respectively. Although recent work enabled in-context surrogate models that do not require re-training, virtually all existing BO methods still require acquisition function maximization to select the next observation, which introduces many knobs to tune, such as Monte Carlo samplers and multi-start optimizers. In this work, we propose a completely in-context, zero-shot solution for BO that does not require surrogate fitting or acquisition function optimization. This is done by using a pre-trained deep generative model to directly sample from the posterior over the optimum point. We show that this process is equivalent to Thompson sampling and demonstrate the capabilities and cost-effectiveness of our foundation model on a suite of real-world benchmarks. We achieve an efficiency gain of more than 35x in terms of wall-clock time when compared with Gaussian process-based BO, enabling efficient parallel and distributed BO, e.g., for high-throughput optimization.
We consider the problem of statistical inference when the data is collected via a Thompson Sampling-type algorithm. While Thompson Sampling (TS) is known to be both asymptotically optimal and empirically effective, its adaptive sampling scheme poses challenges for constructing confidence intervals for model parameters. We propose and analyze a variant of TS, called Stable Thompson Sampling, in which the posterior variance is inflated by a logarithmic factor. We show that this modification leads to asymptotically normal estimates of the arm means, despite the non-i.i.d. nature of the data. Importantly, this statistical benefit comes at a modest cost: the variance inflation increases regret by only a logarithmic factor compared to standard TS. Our results reveal a principled trade-off: by paying a small price in regret, one can enable valid statistical inference for adaptive decision-making algorithms.
Randomized clinical trials often require large patient cohorts before drawing definitive conclusions, yet abundant observational data from parallel studies remains underutilized due to confounding and hidden biases. To bridge this gap, we propose Deconfounded Warm-Start Thompson Sampling (DWTS), a practical approach that leverages a Doubly Debiased LASSO (DDL) procedure to identify a sparse set of reliable measured covariates and combines them with key hidden covariates to form a reduced context. By initializing Thompson Sampling (LinTS) priors with DDL-estimated means and variances on these measured features -- while keeping uninformative priors on hidden features -- DWTS effectively harnesses confounded observational data to kick-start adaptive clinical trials. Evaluated on both a purely synthetic environment and a virtual environment created using real cardiovascular risk dataset, DWTS consistently achieves lower cumulative regret than standard LinTS, showing how offline causal insights from observational data can improve trial efficiency and support more personalized treatment decisions.
We study the problem of selecting a subset from a large action space shared by a family of bandits, with the goal of achieving performance nearly matching that of using the full action space. We assume that similar actions tend to have related payoffs, modeled by a Gaussian process. To exploit this structure, we propose a simple epsilon-net algorithm to select a representative subset. We provide theoretical guarantees for its performance and compare it empirically to Thompson Sampling and Upper Confidence Bound.
This paper presents a concise review of Contextual Multi-Armed Bandit (CMAB) methods and introduces an experimental framework for scalable, interpretable offer selection, addressing the challenge of fast-changing offers. The approach models context at the product category level, allowing offers to span multiple categories and enabling knowledge transfer across similar offers. This improves learning efficiency and generalization in dynamic environments. The framework extends standard CMAB methodology to support multi-category contexts, and achieves scalability through efficient feature engineering and modular design. Advanced features such as MPG (Member Purchase Gap) and MF (Matrix Factorization) capture nuanced user-offer interactions, with implementation in Python for practical deployment. A key contribution is interpretability at scale: logistic regression models yield transparent weight vectors, accessible via a large language model (LLM) interface for real-time, user-level tracking and explanation of evolving preferences. This enables the generation of detailed member profiles and identification of behavioral patterns, supporting personalized offer optimization and enhancing trust in automated decisions. By situating our prototype alongside established paradigms like Generalized Linear Models and Thompson Sampling, we demonstrate its value for both research and real-world CMAB applications.
Recent advances in generative artificial intelligence (GenAI) models have enabled the generation of personalized content that adapts to up-to-date user context. While personalized decision systems are often modeled using bandit formulations, the integration of GenAI introduces new structure into otherwise classical sequential learning problems. In GenAI-powered interventions, the agent selects a query, but the environment experiences a stochastic response drawn from the generative model. Standard bandit methods do not explicitly account for this structure, where actions influence rewards only through stochastic, observed treatments. We introduce generator-mediated bandit-Thompson sampling (GAMBITTS), a bandit approach designed for this action/treatment split, using mobile health interventions with large language model-generated text as a motivating case study. GAMBITTS explicitly models both the treatment and reward generation processes, using information in the delivered treatment to accelerate policy learning relative to standard methods. We establish regret bounds for GAMBITTS by decomposing sources of uncertainty in treatment and reward, identifying conditions where it achieves stronger guarantees than standard bandit approaches. In simulation studies, GAMBITTS consistently outperforms conventional algorithms by leveraging observed treatments to more accurately estimate expected rewards.
In this study, I investigate the dynamic decision problem with a finite parameter space when the functional form of conditional expected rewards is misspecified. Traditional algorithms, such as Thompson Sampling, guarantee neither an $O(e^{-T})$ rate of posterior parameter concentration nor an $O(T^{-1})$ rate of average regret. However, under mild conditions, we can still achieve an exponential convergence rate of the parameter to a pseudo truth set, an extension of the pseudo truth parameter concept introduced by White (1982). I further characterize the necessary conditions for the convergence of the expected posterior within this pseudo-truth set. Simulations demonstrate that while the maximum a posteriori (MAP) estimate of the parameters fails to converge under misspecification, the algorithm's average regret remains relatively robust compared to the correctly specified case. These findings suggest opportunities to design simple yet robust algorithms that achieve desirable outcomes even in the presence of model misspecifications.
Neural Machine Translation (NMT) systems face significant challenges when working with low-resource languages, particularly in domain adaptation tasks. These difficulties arise due to limited training data and suboptimal model generalization, As a result, selecting an optimal model for translation is crucial for achieving strong performance on in-domain data, particularly in scenarios where fine-tuning is not feasible or practical. In this paper, we investigate strategies for selecting the most suitable NMT model for a given domain using bandit-based algorithms, including Upper Confidence Bound, Linear UCB, Neural Linear Bandit, and Thompson Sampling. Our method effectively addresses the resource constraints by facilitating optimal model selection with high confidence. We evaluate the approach across three African languages and domains, demonstrating its robustness and effectiveness in both scenarios where target data is available and where it is absent.
Protein fitness optimization involves finding a protein sequence that maximizes desired quantitative properties in a combinatorially large design space of possible sequences. Recent developments in steering protein generative models (e.g diffusion models, language models) offer a promising approach. However, by and large, past studies have optimized surrogate rewards and/or utilized large amounts of labeled data for steering, making it unclear how well existing methods perform and compare to each other in real-world optimization campaigns where fitness is measured by low-throughput wet-lab assays. In this study, we explore fitness optimization using small amounts (hundreds) of labeled sequence-fitness pairs and comprehensively evaluate strategies such as classifier guidance and posterior sampling for guiding generation from different discrete diffusion models of protein sequences. We also demonstrate how guidance can be integrated into adaptive sequence selection akin to Thompson sampling in Bayesian optimization, showing that plug-and-play guidance strategies offer advantages compared to alternatives such as reinforcement learning with protein language models.
Diffusion models have shown significant potential in generating oracle items that best match user preference with guidance from user historical interaction sequences. However, the quality of guidance is often compromised by unpredictable missing data in observed sequence, leading to suboptimal item generation. Since missing data is uncertain in both occurrence and content, recovering it is impractical and may introduce additional errors. To tackle this challenge, we propose a novel dual-side Thompson sampling-based Diffusion Model (TDM), which simulates extra missing data in the guidance signals and allows diffusion models to handle existing missing data through extrapolation. To preserve user preference evolution in sequences despite extra missing data, we introduce Dual-side Thompson Sampling to implement simulation with two probability models, sampling by exploiting user preference from both item continuity and sequence stability. TDM strategically removes items from sequences based on dual-side Thompson sampling and treats these edited sequences as guidance for diffusion models, enhancing models' robustness to missing data through consistency regularization. Additionally, to enhance the generation efficiency, TDM is implemented under the denoising diffusion implicit models to accelerate the reverse process. Extensive experiments and theoretical analysis validate the effectiveness of TDM in addressing missing data in sequential recommendations.
Stochastic rising rested bandit (SRRB) is a setting where the arms' expected rewards increase as they are pulled. It models scenarios in which the performances of the different options grow as an effect of an underlying learning process (e.g., online model selection). Even if the bandit literature provides specifically crafted algorithms based on upper-confidence bounds for such a setting, no study about Thompson sampling TS-like algorithms has been performed so far. The strong regularity of the expected rewards in the SRRB setting suggests that specific instances may be tackled effectively using adapted and sliding-window TS approaches. This work provides novel regret analyses for such algorithms in SRRBs, highlighting the challenges and providing new technical tools of independent interest. Our results allow us to identify under which assumptions TS-like algorithms succeed in achieving sublinear regret and which properties of the environment govern the complexity of the regret minimization problem when approached with TS. Furthermore, we provide a regret lower bound based on a complexity index we introduce. Finally, we conduct numerical simulations comparing TS-like algorithms with state-of-the-art approaches for SRRBs in synthetic and real-world settings.