Multi-Objective contextual bandits with a dominant objective

In the field of driver assistance systems, many methods have been proposed to estimate the alertness of a driver and detect dangerous driver states. This article addresses a complementary problem – given that there is a method for detecting the alertness of a vehicle's driver, how do we find an effective countermeasure policy to make sure that the driver stays in the alert state, therefore, minimizing safety risks? The article formulates the problem of maintaining driver's state with a help of active countermeasures as a contextual bandit problem. In particular, it proposes a conceptual schema of the approach, defines problem-specific action loss function, context representation, and implements several contextual bandit algorithms. To evaluate the approach the paper introduces a computational model of a driver's alertness, based on fatigue research. The performed evaluation with the computational model has shown that all implemented algorithms just after few simulated trips were able to obtain intervention policies significantly surpassing non-personalized baseline.

The contextual bandit problem has been studied in the recommender system community, but without paying much attention to the contextual aspect of the recommendation. We introduce in this paper an algorithm that tackles this problem by modeling the Mobile Context-Aware Recommender Systems (MCRS) as a contextual bandit algorithm and it is based on dynamic exploration/exploitation. Within a deliberately designed offline simulation framework, we conduct extensive evaluations with real online event log data. The experimental results and detailed analysis demonstrate that our algorithm outperforms surveyed algorithms.

Dynamic Personalized Decision Making Beyond the Super-Extrapolatable and Super-Local Cases Contextual bandit problems model the inherent trade-off between exploration and exploitation in personalized decision making in marketing, healthcare, revenue management, and more. Specifically, the trade-off is characterized by the optimal growth rate of the regret. Intuitively, the optimal rate should depend on how complex the underlying supervised learning problem is, namely, how much can observing reward in one context tell us about mean rewards in another. To formalize this intuitive relationship, Hu, Kallus, and Mao study in “Smooth Contextual Bandits: Bridging the Parametric and Nondifferentiable Regimes” a nonparametric contextual bandit problem in which mean reward functions are β-times differentiable (more generally, Hölder β-smooth). This interpolates between two extremes previously studied in isolation: nondifferentiable bandits (β ≤ 1), with which running separated noncontextual bandits in different context regions achieves rate-optimal regret, and parametric-response bandits (β = ∞), with which rate-optimal regret can be achieved with minimal or no exploration because of infinite extrapolatability across contexts. The authors develop a rate-optimal algorithm that operates neither fully locally nor fully globally, revealing the optimal regret rate in this in-between smooth setting and shedding light on the crucial interplay of functional complexity and regret in dynamic personalized decision making.

Contextual bandit problems are a natural fit for many information retrieval tasks, such as learning to rank, text classification, recommendation, and so on. However, existing learning methods for contextual bandit problems have one of two drawbacks: They either do not explore the space of all possible document rankings (i.e., actions) and, thus, may miss the optimal ranking, or they present suboptimal rankings to a user and, thus, may harm the user experience. We introduce a new learning method for contextual bandit problems, Safe Exploration Algorithm (SEA), which overcomes the above drawbacks. SEA starts by using a baseline (or production) ranking system (i.e., policy), which does not harm the user experience and, thus, is safe to execute but has suboptimal performance and, thus, needs to be improved. Then SEA uses counterfactual learning to learn a new policy based on the behavior of the baseline policy. SEA also uses high-confidence off-policy evaluation to estimate the performance of the newly learned policy. Once the performance of the newly learned policy is at least as good as the performance of the baseline policy, SEA starts using the new policy to execute new actions, allowing it to actively explore favorable regions of the action space. This way, SEA never performs worse than the baseline policy and, thus, does not harm the user experience, while still exploring the action space and, thus, being able to find an optimal policy. Our experiments using text classification and document retrieval confirm the above by comparing SEA (and a boundless variant called BSEA) to online and offline learning methods for contextual bandit problems.

Personalization has become a focal point of modern revenue management. However, it is often the case that minimal data are available to appropriately make suggestions tailored to each customer. This has led to many products making use of reinforcement learning-based algorithms to explore sets of offerings to find the best suggestions to improve conversion and revenue. Arguably the most popular of these algorithms are built on the foundation of the multi-arm bandit framework, which has shown great success across a variety of use cases. A general multi-arm bandit algorithm aims to trade-off adaptively exploring available, but under observed, recommendations, with the current known best offering. While much success has been achieved with these relatively understandable procedures, much of the airline industry is losing out on better personalized offers by ignoring the context of the transaction, as is the case in the traditional multi-arm bandit setup. Here, we explore a popular exploration heuristic, Thompson sampling, and note implementation details for multi-arm and contextual bandit variants. While the contextual bandit requires greater computational and technical complexity to include contextual features in the decision process, we illustrate the value it brings by the improvement in overall expected

Personalized medicine and adaptive clinical trials aim to match optimal treatment plans to individual patients while validating new therapies. Traditional fixed-design trials have limitations, including resource wastage and scarce data. Artificial intelligence has led to the development of dynamic decision algorithms like the Multi-Armed Bandit (MAB) algorithm, which balances exploration and exploitation in treatment allocation. Researchers aim to integrate MAB with clinical needs while adhering to ethical guidelines. This paper discusses the use of Multi-armed Bandit (MAB) algorithms in medicine, highlighting their potential for optimizing treatment allocation and improving patient outcomes, despite ethical constraints and limited data, and their application in contextual and reinforcement learning settings. This research highlight key clinical applications such as adaptive dose-finding, personalized treatment selection, and digital health interventions, supported by both trial-based data and large-scale public datasets like MIMIC-III. Simulation studies are also discussed as a necessary complement to real-world data, facilitating algorithm validation under ethical and logistical constraints. Comparative evaluation of algorithms demonstrates that Bayesian methods, particularly Thompson Sampling and contextual bandits, often provide a more robust balance between efficiency and safety. However, challenges remain in scalability, interpretability, and regulatory acceptance. This research conclude by identifying promising directions for future research, including the integration of deep reinforcement learning and causal inference, which may further enhance the role of MABs in advancing personalized medicine and adaptive clinical trial design.

This paper presents a new contextual bandit algorithm, NeuralBandit, which does not need hypothesis on stationarity of contexts and rewards. Several neural networks are trained to modelize the value of rewards knowing the context. Two variants, based on multi-experts approach, are proposed to choose online the parameters of multi-layer perceptrons. The proposed algorithms are successfully tested on a large dataset with and without stationarity of rewards.KeywordsConcept DriftBandit ProblemForest Cover TypeMultiarmed BanditNeural Network CommitteeThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

The contextual bandit problem, where agents arrive sequentially with personal contexts and the system adapts its arm allocation decisions accordingly, has recently garnered increasing attention for enabling more personalized outcomes. However, in many healthcare and recommendation applications, agents have private profiles and may misreport their contexts to gain from the system. For example, in adaptive clinical trials, where hospitals sequentially recruit volunteers to test multiple new treatments and adjust plans based on volunteers' reported profiles such as symptoms and interim data, participants may misreport severe side effects like allergy and nausea to avoid perceived suboptimal treatments. We are the first to study this issue of private context misreporting in a stochastic contextual bandit game between the system and non-repeated agents. We show that traditional low-regret algorithms, such as UCB family algorithms and Thompson sampling, fail to ensure truthful reporting and can result in linear regret in the worst case, while traditional truthful algorithms like explore-then-commit (ETC) and ε-greedy algorithm incur sublinear but high regret. We propose a mechanism that uses a linear program to ensure truthfulness while minimizing deviation from Thompson sampling, yielding an O(ln T) frequentist regret. Our numerical experiments further demonstrate strong performance in multiple contexts and across other distribution families.

Contextual multi-armed bandit (MAB) is a classic online learning problem, where a learner/agent selects actions (i.e., arms) given contextual information and discovers optimal actions based on reward feedback. Applications of contextual bandit have been increasingly expanding, including advertisement, personalization, resource allocation in wireless networks, among others. Nonetheless, the reward feedback is delayed in many applications (e.g., a user may only provide service ratings after a period of time), creating challenges for contextual bandits. In this paper, we address delayed feedback in contextual bandits by using semi-supervised learning — incorporate estimates of delayed rewards to improve the estimation of future rewards. Concretely, the reward feedback for an arm selected at the beginning of a round is only observed by the agent/learner with some observation noise and provided to the agent after some a priori unknown but bounded delays. Motivated by semi-supervised learning that produces pseudo labels for unlabeled data to further improve the model performance, we generate fictitious estimates of rewards that are delayed and have yet to arrive based on already-learnt reward functions. Thus, by combining semi-supervised learning with online contextual bandit learning, we propose a novel extension and design two algorithms, which estimate the values for currently unavailable reward feedbacks to minimize the maximum estimation error and average estimation error, respectively.

Contextual bandit algorithms are ubiquitous tools for active sequential experimentation in healthcare and the tech industry. They involve online learning algorithms that adaptively learn policies over time to map observed contexts X t to actions A t in an attempt to maximize stochastic rewards R t . This adaptivity raises interesting but hard statistical inference questions, especially counterfactual ones: for example, it is often of interest to estimate the properties of a hypothetical policy that is different from the logging policy that was used to collect the data—a problem known as “off-policy evaluation” (OPE). Using modern martingale techniques, we present a comprehensive framework for OPE inference that relaxes unnecessary conditions made in some past works (such as performing inference at prespecified sample sizes, uniformly bounded importance weights, constant logging policies, and constant policy values, among others), significantly improving on them both theoretically and empirically. Importantly, our methods can be employed while the original experiment is still running (that is, not necessarily post hoc), when the logging policy may be itself changing (due to learning), and even if the context distributions are a highly dependent time series (such as if they are drifting over time). More concretely, we derive confidence sequences for various functionals of interest in OPE. These include doubly robust ones for time-varying off-policy mean reward values, but also confidence bands for the entire cumulative distribution function of the off-policy reward distribution. All of our methods (a) are valid at arbitrary stopping times; (b) only make nonparametric assumptions; (c) do not require importance weights to be uniformly bounded, and if they are, we do not need to know these bounds; and (d) adapt to the empirical variance of our estimators. In summary, our methods enable anytime-valid off-policy inference using adaptively collected contextual bandit data.

Reinforcement learning algorithms play an important role in modern day and have been applied to many domains. For example, personalized recommendations problem can be modelled as a contextual multi-armed bandit problem in reinforcement learning. In this paper, we propose a contextual bandit algorithm which is based on Contexts and the Chosen Number of Arm with Minimal Estimation, namely Con-CNAME in short. The continuous exploration and context used in our algorithm can address the cold start problem in recommender systems. Furthermore, the Con-CNAME algorithm can still make recommendations under the emergency circumstances where contexts are unavailable suddenly. In the experimental evaluation, the reference range of key parameters and the stability of Con-CNAME are discussed in detail. In addition, the performance of Con-CNAME is compared with some classic algorithms. Experimental results show that our algorithm outperforms several bandit algorithms.

District heating networks make up an important public energy service, in which leakage is the main problem affecting the safety of pipeline network operation. This paper proposes a Leakage Fault Detection (LFD) method based on the Linear Upper Confidence Bound (LinUCB) which is used for arm selection in the Contextual Bandit (CB) algorithm. With data collected from end-users’ pressure and flow information in the simulation model, the LinUCB method is adopted to locate the leakage faults. Firstly, we use a hydraulic simulation model to simulate all failure conditions that can occur in the network, and these change rate vectors of observed data form a dataset. Secondly, the LinUCB method is used to train an agent for the arm selection, and the outcome of arm selection is the leaking pipe label. Thirdly, the experiment results show that this method can detect the leaking pipe accurately and effectively. Furthermore, it allows operators to evaluate the system performance, supports troubleshooting of decision mechanisms, and provides guidance in the arrangement of maintenance.

Recommender systems play a critical role in e-commerce, streaming media, and other fields, where user engagement is vital to platform success. These systems, however, face challenges like the cold start problem, which limits recommendation accuracy for new users and items due to sparse data. Although multi-armed bandit (MAB) algorithms are considered effective for balancing exploration and exploitation, existing studies lack simulations of algorithm performance in scenarios where recommender systems indirectly influence user choices, limiting insights for performance improvements in real-world applications. This study explores the performance of key MAB algorithmsExplore-Then-Commit (ETC), Upper Confidence Bound (UCB), and Thompson Sampling (TS)under a novel framework that models user behavior as an unreliable intermediate using Markov state transitions. This Markov perturbation-based approach provides a foundation for indirectly influencing user decisions. Two algorithmic frameworks are examined: one considering only the recommendation system's influence on user choices, and another also considering user feedback as a dynamic influence on state transition probabilities. Results show that, while considering user feedback increases adaptability, it also causes greater volatility. TS achieves the lowest cumulative regret with strong stability, while Asymptotic-UCB achieves a better balance between exploration and exploitation than standard UCB, with consistent adaptability under varying feedback. In contrast, ETC demonstrates significant volatility, highlighting its limitations in feedback-driven scenarios. These findings provide valuable insights for enhancing MAB algorithms in real-world cold start recommendation applications, especially where systems influence user choices indirectly.

Collaborative bandit learning, i.e., bandit algorithms that utilize collaborative filtering techniques to improve sample efficiency in online interactive recommendation, has attracted much research attention as it enjoys the best of both worlds. However, all existing collaborative bandit learning solutions impose a stationary assumption about the environment, i.e., both user preferences and the dependency among users are assumed static over time. Unfortunately, this assumption hardly holds in practice due to users' ever-changing interests and dependency relations, which inevitably costs a recommender system sub-optimal performance in practice. In this work, we develop a collaborative dynamic bandit solution to handle a changing environment for recommendation. We explicitly model the underlying changes in both user preferences and their dependency relation as a stochastic process. Individual user's preference is modeled by a mixture of globally shared contextual bandit models with a Dirichlet process prior. Collaboration among users is thus achieved via Bayesian inference over the global bandit models. To balance exploitation and exploration during the interactions, Thompson sampling is used for both model selection and arm selection. Our solution is proved to maintain a standard ~O(√T) Bayesian regret in this challenging environment. Extensive empirical evaluations on both synthetic and real-world datasets further confirmed the necessity of modeling a changing environment and our algorithm's practical advantages against several state-of-the-art online learning solutions.

In sequential decision problems in an unknown environment, the decision maker often faces a dilemma over whether to explore to discover more about the environment, or to exploit current knowledge. We address the exploration-exploitation dilemma in a general setting encompassing both standard and contextualised bandit problems. The contextual bandit problem has recently resurfaced in attempts to maximise click-through rates in web based applications, a task with significant commercial interest. In this article we consider an approach of Thompson (1933) which makes use of samples from the posterior distributions for the instantaneous value of each action. We extend the approach by introducing a new algorithm, Optimistic Bayesian Sampling (OBS), in which the probability of playing an action increases with the uncertainty in the estimate of the action value. This results in better directed exploratory behaviour. We prove that, under unrestrictive assumptions, both approaches result in optimal behaviour with respect to the average reward criterion of Yang and Zhu (2002). We implement OBS and measure its performance in simulated Bernoulli bandit and linear regression domains, and also when tested with the task of personalised news article recommendation on a Yahoo! Front Page Today Module data set. We find that OBS performs competitively when compared to recently proposed benchmark algorithms and outperforms Thompson's method throughout.