Carousels have become the de-facto standard user interface in online services. However, there is a lack of research in carousels, particularly examining how recommender systems may be designed differently than the traditional single-list interfaces. One of the key elements for understanding how to design a system for a particular interface is understanding how users browse. For carousels, users may browse in a number of different ways due to the added complexity of multiple topic defined-lists and swiping to see more items.

We study estimator selection and hyper-parameter tuning in off-policy evaluation. Although cross-validation is the most popular method for model selection in supervised learning, off-policy evaluation relies mostly on theory, which provides only limited guidance to practitioners. We show how to use cross-validation for off-policy evaluation. This challenges a popular belief that cross-validation in off-policy evaluation is not feasible. We evaluate our method empirically and show that it addresses a variety of use cases.

We consider the finite-horizon offline reinforcement learning (RL) setting, and are motivated by the challenge of learning the policy at any step h in dynamic programming (DP) algorithms. To learn this, it is sufficient to evaluate the treatment effect of deviating from the behavioral policy at step h after having optimized the policy for all future steps. Since the policy at any step can affect next-state distributions, the related distributional shift challenges can make this problem far more statistically hard than estimating such treatment effects in the stochastic contextual bandit setting. However, the hardness of many real-world RL instances lies between the two regimes. We develop a flexible and general method called selective uncertainty propagation for confidence interval construction that adapts to the hardness of the associated distribution shift challenges. We show benefits of our approach on toy environments and demonstrate the benefits of these techniques for offline policy learning.

Yu Xia, Subhojyoti Mukherjee, Zhouhang Xie, Junda Wu, Xintong Li, Ryan Aponte, Hanjia Lyu, Joe Barrow, Hongjie Chen, Franck Dernoncourt, Branislav Kveton, Tong Yu, Ruiyi Zhang, Jiuxiang Gu, Nesreen K. Ahmed, Yu Wang, Xiang Chen, Hanieh Deilamsalehy, Sungchul Kim, Zhengmian Hu, Yue Zhao, Nedim Lipka, Seunghyun Yoon, Ting-Hao Kenneth Huang, Zichao Wang, Puneet Mathur, Soumyabrata Pal, Koyel Mukherjee, Zhehao Zhang, Namyong Park, Thien Huu Nguyen, Jiebo Luo, Ryan A. Rossi, Julian McAuley. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Offline evaluation of LLMs is crucial in understanding their capacities, though current methods remain underexplored in existing research. In this work, we focus on the offline evaluation of the chain-of-thought capabilities and show how to optimize LLMs based on the proposed evaluation method. To enable offline feedback with rich knowledge and reasoning paths, we use knowledge graphs (KGs) (e.g., Wikidata5M) to provide feedback on the generated chain of thoughts. Due to the heterogeneity between LLM reasoning and KG structures, direct interaction and feedback from knowledge graphs on LLM behavior are challenging, as they require accurate entity linking and grounding of LLM-generated chains of thought in the KG. To address the above challenge, we propose an offline chain-of-thought evaluation framework, OCEAN, which models chain-of-thought reasoning in LLMs as a Markov Decision Process (MDP), and evaluate the policy’s alignment with KG preference modeling. To overcome the reasoning heterogeneity and grounding problems, we leverage on-policy KG exploration and reinforcement learning to model a KG policy that generates token-level likelihood distributions for LLM-generated chain-of-thought reasoning paths, simulating KG reasoning preference. Then we incorporate the knowledge-graph feedback on the validity and alignment of the generated reasoning paths into inverse propensity scores and propose KG-IPS estimator. Theoretically, we prove the unbiasedness of the proposed KG-IPS estimator and provide a lower bound on its variance. With the off-policy evaluated value function, we can directly enable off-policy optimization to further enhance chain-of-thought alignment. Our empirical study shows that OCEAN can be efficiently optimized for generating chain-of-thought reasoning paths with higher estimated values without affecting LLMs’ general abilities in downstream tasks or their internal knowledge.

Supervised fine-tuning (SFT) is the most common way of adapting large language models (LLMs) to a new domain. In this paper, we improve the efficiency of SFT by selecting an informative subset of training examples. Specifically, for a fixed budget of training examples, which determines the computational cost of fine-tuning, we select those that maximize information gain, as measured by the Fisher information matrix of the SFT objective. We approximate it efficiently by linearization at the last layer of the LLM. Our approach is computationally efficient, analyzable, and performs well empirically. We demonstrate this on several problems, with both quantitative results and LLM-as-a-judge evaluations.

We study learning human preferences from limited comparison feedback, a core machine learning problem that is at the center of reinforcement learning from human feedback (RLHF). We formulate the problem as learning a Plackett-Luce (PL) model from a limited number of $K$-subset comparisons over a universe of $N$ items, where typically $K \ll N$. Our objective is to select the $K$-subsets such that all items can be ranked with minimal mistakes within the budget. We solve the problem using the D-optimal design, which minimizes the worst-case ranking loss under the estimated PL model. All known algorithms for this problem are computationally infeasible in our setting because we consider exponentially many subsets in $K$. To address this challenge, we propose a randomized Frank-Wolfe algorithm with memoization and sparse updates that has a low $O(N^2 + K^2)$ per-iteration complexity. We analyze it and demonstrate its empirical superiority on synthetic and open-source NLP datasets.

Offline reinforcement learning (RL) is a variant of RL where the policy is learned from a previously collected dataset of trajectories and rewards. In our work, we propose a practical approach to offline RL with large language models (LLMs). We recast the problem as reward-weighted fine-tuning, which can be solved using similar techniques to supervised fine-tuning (SFT). To showcase the value of our approach, we apply it to learning short-horizon question-answering policies of a fixed length, where the agent reasons about potential answers or asks clarifying questions. Our work stands in a stark contrast to state-of-the-art methods in this domain, based on SFT and direct preference optimization, which have additional hyper-parameters and do not directly optimize for rewards. We compare to them empirically, and report major gains in both optimized rewards and language quality.

Carousel interfaces are widely used in e-commerce and streaming services, but little research has been devoted to them. Previous studies of interfaces for presenting search and recommendation results have focused on single ranked lists, but it appears their results cannot be extrapolated to carousels due to the added complexity. Eye tracking is a highly informative approach to understanding how users click, yet there are no eye tracking studies concerning carousels. There are very few interaction datasets on recommenders with carousel interfaces and none that contain gaze data. We introduce the RecGaze dataset: the first comprehensive feedback dataset on carousels that includes eye tracking results, clicks, cursor movements, and selection explanations. The dataset comprises of interactions from 3 movie selection tasks with 40 different carousel interfaces per user. In total, 87 users and 3,477 interactions are logged. In addition to the dataset, its description and possible use cases, we provide results of a survey on carousel design and the first analysis of gaze data on carousels, which reveals a golden triangle or F-pattern browsing behavior. Our work seeks to advance the field of carousel interfaces by providing the first dataset with eye tracking results on carousels. In this manner, we provide and encourage an empirical understanding of interactions with carousel interfaces, for building better recommender systems through gaze information, and also encourage the development of gaze-based recommenders.

Multi-objective optimization is a class of optimization problems with multiple conflicting objectives. We study offline optimization of multi-objective policies from data collected by a previously deployed policy. We propose a pessimistic estimator for policy values that can be easily plugged into existing formulas for hypervolume computation and optimized. The estimator is based on inverse propensity scores (IPS), and improves upon a naive IPS estimator in both theory and experiments. Our analysis is general, and applies beyond our IPS estimators and methods for optimizing them.

Off-policy learning is a framework for optimizing policies without deploying them, using data collected by another policy. In recommender systems, this is especially challenging due to the imbalance in logged data: some items are recommended and thus logged more frequently than others. This is further perpetuated when recommending a list of items, as the action space is combinatorial. To address this challenge, we study pessimistic off-policy optimization for learning to rank. The key idea is to compute lower confidence bounds on parameters of click models and then return the list with the highest pessimistic estimate of its value. This approach is computationally efficient, and we analyze it. We study its Bayesian and frequentist variants and overcome the limitation of unknown prior by incorporating empirical Bayes. To show the empirical effectiveness of our approach, we compare it to off-policy optimizers that use inverse propensity scores or neglect uncertainty. Our approach outperforms all baselines and is both robust and general.

Most bandit algorithms assume that the reward variances or their upper bounds are known, and that they are the same for all arms. This naturally leads to suboptimal performance and higher regret due to variance overestimation. On the other hand, underestimated reward variances may lead to linear regret due to committing early to a suboptimal arm. This motivated prior works on variance-adaptive frequentist algorithms, which have strong instance-dependent regret bounds but cannot incorporate prior knowledge on reward variances. We lay foundations for the Bayesian setting, which incorporates prior knowledge. This results in lower regret in practice, since the prior is used in the algorithm design, and also improved regret guarantees. Specifically, we study Gaussian bandits with \emph{unknown heterogeneous reward variances} and develop a Thompson sampling algorithm with prior-dependent Bayes regret bounds. We achieve lower regret with lower reward variances and more informative priors on them, which is precisely why we pay only for what is uncertain. This is the first such result in the bandit literature. Finally, we corroborate our theory with experiments, which demonstrate the benefit of our variance-adaptive Bayesian algorithm over prior frequentist works. We also show that our approach is robust to model misspecification and can be applied with estimated priors.

Following the introduction of Adam, several novel adaptive optimizers for deep learning have been proposed. These optimizers typically excel in some tasks but may not outperform Adam uniformly across all tasks. In this work, we introduce Meta-Adaptive Optimizers (MADA), a unified optimizer framework that can generalize several known optimizers and dynamically learn the most suitable one during training. The key idea in MADA is to parameterize the space of optimizers and dynamically search through it using hyper-gradient descent during training. We empirically compare MADA to other popular optimizers on vision and language tasks, and find that MADA consistently outperforms Adam and other popular optimizers, and is robust against sub-optimally tuned hyper-parameters. MADA achieves a greater validation performance improvement over Adam compared to other popular optimizers during GPT-2 training and fine-tuning. We also propose AVGrad, a modification of AMSGrad that replaces the maximum operator with averaging, which is more suitable for hyper-gradient optimization. Finally, we provide a convergence analysis to show that parameterized interpolations of optimizers can improve their error bounds (up to constants), hinting at an advantage for meta-optimizers.

Posterior sampling in contextual bandits with a Gaussian prior can be implemented exactly or approximately using the Laplace approximation. The Gaussian prior is computationally efficient but it cannot describe complex distributions. In this work, we propose approximate posterior sampling algorithms for contextual bandits with a diffusion model prior. The key idea is to sample from a chain of approximate conditional posteriors, one for each stage of the reverse diffusion process, which are obtained by the Laplace approximation. Our approximations are motivated by posterior sampling with a Gaussian prior, and inherit its simplicity and efficiency. They are asymptotically consistent and perform well empirically on a variety of contextual bandit problems.

Learning of preference models from human feedback has been central to recent advances in artificial intelligence. Motivated by the cost of obtaining high-quality human annotations, we study efficient human preference elicitation for learning preference models. The key idea in our work is to generalize optimal designs, an approach to computing optimal information-gathering policies, to lists of items that represent potential questions with answers. The policy is a distribution over the lists and we elicit preferences from them proportionally to their probabilities. To show the generality of our ideas, we study both absolute and ranking feedback models on items in the list. We design efficient algorithms for both and analyze them. Finally, we demonstrate that our algorithms are practical by evaluating them on existing question-answering problems.

Recent studies on pre-trained vision/language models have demonstrated the practical benefit of a new, promising solution-building paradigm in AI where models can be pre-trained on broad data describing a generic task space and then adapted successfully to solve a wide range of downstream tasks, even when training data is severely limited (e.g., in zero- or few-shot learning scenarios). Inspired by such progress, we investigate in this paper the possibilities and challenges of adapting such a paradigm to the context of recommender systems, which is less investigated from the perspective of pre-trained model. In particular, we propose to develop a generic recommender that captures universal interaction patterns by training on generic user-item interaction data extracted from different domains, which can then be fast adapted to improve few-shot learning performance in unseen new domains (with limited data).

The unique capabilities of Large Language Models (LLMs), such as the natural language text generation ability, position them as strong candidates for providing explanation for recommendations. However, despite the size of the LLM, most existing models struggle to produce zero-shot explanations reliably. To address this issue, we propose a framework called Logic-Scaffolding, that combines the ideas of aspect-based explanation and chain-of-thought prompting to generate explanations through intermediate reasoning steps. In this paper, we share our experience in building the framework and present an interactive demonstration for exploring our results.

Offline data-driven evaluation is considered a low-cost and more accessible alternative to the online empirical method of assessing the quality of recommender systems. Despite their popularity and effectiveness, most data-driven approaches are unsuitable for evaluating interactive recommender systems. In this article, we attempt to address this issue by simulating the user interactions with the system as a part of the evaluation process. Particularly, we demonstrate that simulated users find their desired item more efficiently when recommendations are presented as a list of carousels compared to a simple ranked list.

We develop a meta-learning framework for simple regret minimization in bandits. In this framework, a learning agent interacts with a sequence of bandit tasks, which are sampled i.i.d. from an unknown prior distribution, and learns its meta-parameters to perform better on future tasks. We propose the first Bayesian and frequentist meta-learning algorithms for this setting. The Bayesian algorithm has access to a prior distribution over the meta-parameters and its meta simple regret over m bandit tasks with horizon n is mere O(m / √n). On the other hand, the meta simple regret of the frequentist algorithm is O(n√m + m/ √n). While its regret is worse, the frequentist algorithm is more general because it does not need a prior distribution over the meta-parameters. It can also be analyzed in more settings. We instantiate our algorithms for several classes of bandit problems. Our algorithms are general and we complement our theory by evaluating them empirically in several environments.

A contextual bandit is a popular framework for online learning to act under uncertainty. In practice, the number of actions is huge and their expected rewards are correlated. In this work, we introduce a general framework for capturing such correlations through a mixed-effect model where actions are related through multiple shared effect parameters. To explore efficiently using this structure, we propose Mixed-Effect Thompson Sampling (meTS) and bound its Bayes regret. The regret bound has two terms, one for learning the action parameters and the other for learning the shared effect parameters. The terms reflect the structure of our model and the quality of priors. Our theoretical findings are validated empirically using both synthetic and real-world problems. We also propose numerous extensions of practical interest. While they do not come with guarantees, they perform well empirically and show the generality of the proposed framework.

Bandit algorithms arose as a standard approach to learning better models online. As they become more popular, they are increasingly deployed in complex machine learning pipelines, where their actions can be overwritten. For example, in ranking problems, a list of recommended items can be modified by a downstream algorithm to increase diversity. This may break the classic bandit algorithms and lead to linear regret. Specifically, if the proposed action is not taken, uncertainty in its estimated mean reward may not get reduced. In this work, we study this setting and call it non-compliant bandits; as the agent tries to learn rewarding actions that comply with a downstream task. We propose two algorithms, compliant contextual UCB (CompUCB) and Thompson sampling (CompTS), which learn separate reward and compliance models. The compliance model allows the agent to avoid non-compliant actions. We derive a sublinear regret bound for CompUCB. We also conduct experiments that compare our algorithms to classic bandit baselines. The experiments show failures of the baselines and that we mitigate them by learning compliance models.

Many practical problems involve solving similar tasks. In recommender systems, the tasks can be users with similar preferences; in search engines, the tasks can be items with similar affinities. To learn statistically efficiently, the tasks can be organized in a hierarchy, where the task affinity is captured using an unknown latent parameter. We study the problem of off-policy learning for similar tasks from logged bandit feedback. To solve the problem, we propose a hierarchical off-policy optimization algorithm HierOPO. The key idea is to estimate the task parameters using the hierarchy and then act pessimistically with respect to them. To analyze the algorithm, we develop novel Bayesian error bounds. Our bounds are the first in off-policy learning that improve with a more informative prior and capture statistical gains due to hierarchical models. Therefore, they are of a general interest. HierOPO also performs well in practice. Our experiments demonstrate the benefits of using the hierarchy over solving each task independently.

In this work, we initiate the idea of using denoising diffusion models to learn priors for online decision making problems. We specifically focus on bandit meta-learning, aiming to learn a policy that performs well across bandit tasks of a same class. To this end, we train a diffusion model that learns the underlying task distribution and combine Thompson sampling with the learned prior to deal with new tasks at test time. Our posterior sampling algorithm carefully balances between the learned prior and the noisy observations that come from the learner's interaction with the environment. To capture realistic bandit scenarios, we propose a novel diffusion model training procedure that trains from incomplete and noisy data, which could be of independent interest. Finally, our extensive experiments clearly demonstrate the potential of the proposed approach.

Despite the great interest in the bandit problem, designing efficient algorithms for complex models remains challenging, as there is typically no analytical way to quantify uncertainty. In this paper, we propose Multiplier Bootstrap-based Exploration (MBE), a novel exploration strategy that is applicable to any reward model amenable to weighted loss minimization. We prove both instance-dependent and instance-independent rate-optimal regret bounds for MBE in sub-Gaussian multi-armed bandits. With extensive simulation and real-data experiments, we show the generality and adaptivity of MBE.

Modern recommender systems usually include separate recommendation carousels such as ‘trending now’ to list trending items and further boost their popularity, thereby attracting active users. Though widely useful, such ‘trending now’ carousels typically generate item lists based on simple heuristics, e.g., the number of interactions within a time interval, and therefore still leave much room for improvement. This paper aims to systematically study this under-explored but important problem from the new perspective of time series forecasting. We first provide a set of rigorous definitions related to item trendiness and formulate the trend recommendation task as a one-step time series forecasting problem. We then propose a deep latent variable model, dubbed Trend Recommender (TrendRec), to forecast items’ future trends and generate trending item lists. Furthermore, we design associated evaluation protocols for trend recommendation. Experiments on real-world datasets from various domains show that our TrendRec significantly outperforms the baselines, verifying our model’s effectiveness.

We study the problem of best-arm identification (BAI) in the fixed-budget setting with heterogeneous reward variances. We propose two variance-adaptive BAI algorithms for this setting: SHVar for known reward variances and SHAdaVar for unknown reward variances. Our algorithms rely on non-uniform budget allocations among the arms where the arms with higher reward variances are pulled more often than those with lower variances. The main algorithmic novelty is in the design of SHAdaVar, which allocates budget greedily based on overestimating the unknown reward variances. We bound probabilities of misidentifying the best arms in both SHVar and SHAdaVar. Our analyses rely on novel lower bounds on the number of pulls of an arm that do not require closed-form solutions to the budget allocation problem. Since one of our budget allocation problems is analogous to the optimal experiment design with unknown variances, we believe that our results are of a broad interest. Our experiments validate our theory, and show that SHVar and SHAdaVar outperform algorithms from prior works with analytical guarantees.

Meta-, multi-task, and federated learning can be all viewed as solving similar tasks, drawn from a distribution that reflects task similarities. We provide a unified view of all these problems, as learning to act in a hierarchical Bayesian bandit. We propose and analyze a natural hierarchical Thompson sampling algorithm (HierTS) for this class of problems. Our regret bounds hold for many variants of the problems, including when the tasks are solved sequentially or in parallel; and show that the regret decreases with a more informative prior. Our proofs rely on a novel total variance decomposition that can be applied beyond our models. Our theory is complemented by experiments, which show that the hierarchy helps with knowledge sharing among the tasks. This confirms that hierarchical Bayesian bandits are a universal and statistically-efficient tool for learning to act with similar bandit tasks.

We study Thompson sampling (TS) in online decision making, where the uncertain environment is sampled from a mixture distribution. This is relevant in multi-task learning, where a learning agent faces different classes of problems. We incorporate this structure in a natural way by initializing TS with a mixture prior, and call the resulting algorithm MixTS. To analyze MixTS, we develop a novel and general proof technique for analyzing the concentration of mixture distributions. We use it to derive Bayes regret bounds for MixTS in both linear bandits and finite-horizon reinforcement learning (RL). Our regret bounds reflect the structure of the mixture prior, and depend on the number of mixture components and their width. We demonstrate the empirical effectiveness of MixTS in synthetic and real-world experiments.

This paper addresses the cold-start problem in online learning to rank (OLTR). We show both theoretically and empirically that priors improve the quality of ranked lists presented to users interactively based on user feedback. These priors can come in the form of unbiased estimates of the relevance of the ranked items, or more practically, can be obtained from offline-learned models. Our experiments show the effectiveness of priors in improving the short-term regret of tabular OLTR algorithms, based on Thompson sampling and BayesUCB.

Motivated by practical needs in online experimentation and off-policy learning, we study the problem of safe optimal design, where we develop a data logging policy that efficiently explores while achieving competitive rewards with a baseline production policy. We first show, perhaps surprisingly, that a common practice of mixing the production policy with uniform exploration, despite being safe, is sub-optimal in maximizing information gain. Then we propose a safe optimal logging policy for the case when no side information about the actions’ expected rewards is available. We improve upon this design by considering side information and also extend both approaches to a large number of actions with a linear reward model. We analyze how our data logging policies impact errors in off-policy learning. Finally, we empirically validate the benefit of our designs by conducting extensive experiments.

This paper studies regret minimization in a multi-armed bandit. It is well known that side information, such as the prior distribution of arm means in Thompson sampling, can improve the statistical efficiency of the bandit algorithm. While the prior is a blessing when correctly specified, it is a curse when misspecified. To address this issue, we introduce the assumption of a random-effect model to bandits. In this model, the mean arm rewards are drawn independently from an unknown distribution, which we estimate. We derive a random-effect estimator of the arm means, analyze its uncertainty, and design a UCB algorithm ReUCB that uses it. We analyze ReUCB and derive an upper bound on its n-round Bayes regret, which improves upon not using the random-effect structure. Our experiments show that ReUCB can outperform Thompson sampling, without knowing the prior distribution of arm means.

Mean rewards of actions are often correlated. The form of these correlations may be complex and unknown a priori, such as the preferences of users for recommended products and their categories. To maximize statistical efficiency, it is important to leverage these correlations when learning. We formulate a bandit variant of this problem where the correlations of mean action rewards are represented by a hierarchical Bayesian model with latent variables. Since the hierarchy can have multiple layers, we call it deep. We propose a hierarchical Thompson sampling algorithm (HierTS) for this problem and show how to implement it efficiently for Gaussian hierarchies. The efficient implementation is possible due to a novel exact hierarchical representation of the posterior, which itself is of independent interest. We use this exact posterior to analyze the Bayes regret of HierTS. Our regret bounds reflect the structure of the problem, that the regret decreases with more informative priors, and can be recast to highlight reduced dependence on the number of actions. We confirm these theoretical findings empirically, in both synthetic and real-world experiments.

High-quality data plays a central role in ensuring the accuracy of policy evaluation. This paper initiates the study of efficient and safe data collection for bandit policy evaluation. We formulate the problem and investigate its several representative variants. For each variant, we analyze its statistical properties, derive the corresponding exploration policy, and design an efficient algorithm for computing it. Both theoretical analysis and experiments support the usefulness of the proposed methods.

Best-arm identification (BAI) in a fixed-budget setting is a bandit problem where the learning agent maximizes the probability of identifying the optimal (best) arm after a fixed number of observations. Most works on this topic study unstructured problems with a small number of arms, which limits their applicability. We propose a general tractable algorithm that incorporates the structure, by successively eliminating suboptimal arms based on their mean reward estimates from a joint generalization model. We analyze our algorithm in linear and generalized linear models (GLMs), and propose a practical implementation based on a G-optimal design. In linear models, our algorithm has competitive error guarantees to prior works and performs at least as well empirically. In GLMs, this is the first practical algorithm with analysis for fixed-budget BAI.

Most real-world optimization problems have multiple objectives. A system designer needs to find a policy that trades off these objectives to reach a desired operating point. This problem has been studied extensively in the setting of known objective functions. However, we consider a more practical but challenging setting of unknown objective functions. In industry, optimization under this setting is mostly approached with online A/B testing, which is often costly and inefficient. As an alternative, we propose Interactive Multi-Objective Off-policy Optimization (IMO^3). The key idea of IMO^3 is to interact with a system designer using policies evaluated in an off-policy fashion to uncover which policy maximizes her unknown utility function. We theoretically show that IMO^3 identifies a near-optimal policy with high probability, depending on the amount of designer's feedback and training data for off-policy estimation. We demonstrate its effectiveness empirically on several multi-objective optimization problems.

We introduce a new multi-armed bandit model where the reward is a sum of multiple random variables, and each action only alters the distributions of some of these variables. Upon taking an action, the agent observes the realizations of all variables. This model is motivated by marketing campaigns and recommender systems, where the variables represent outcomes on individual customers, such as clicks. We propose UCB-style algorithms that estimate the uplifts of the actions over a baseline. We study multiple variants of the problem, including when the baseline and affected variables are unknown, and prove sublinear regret bounds for all of these. In addition, we provide regret lower bounds that justify the necessity of our modeling assumptions. Experiments on synthetic and real-world datasets demonstrate the benefit of methods that estimate the uplifts over policies that do not use this structure.

Off-policy learning is a framework for evaluating and optimizing policies without deploying them, from data collected by another policy. Real-world environments are typically non-stationary and the offline learned policies should adapt to these changes. To address this challenge, we study the novel problem of off-policy optimization in piecewise-stationary contextual bandits. Our proposed solution has two phases. In the offline learning phase, we partition logged data into categorical latent states and learn a near-optimal sub-policy for each state. In the online deployment phase, we adaptively switch between the learned sub-policies based on their performance. This approach is practical and analyzable, and we provide guarantees on both the quality of off-policy optimization and the regret during online deployment. To show the effectiveness of our approach, we compare it to state-of-the-art baselines on both synthetic and real-world datasets. Our approach outperforms methods that act only on observed context.

Efficient exploration in bandits is a fundamental online learning problem. We propose a variant of Thompson sampling that learns to explore better as it interacts with bandit instances drawn from an unknown prior. The algorithm meta-learns the prior and thus we call it MetaTS. We propose several efficient implementations of MetaTS and analyze it in Gaussian bandits. Our analysis shows the benefit of meta-learning and is of a broader interest, because we derive a novel prior-dependent Bayes regret bound for Thompson sampling. Our theory is complemented by empirical evaluation, which shows that MetaTS quickly adapts to the unknown prior.

We propose AdaTS, a Thompson sampling algorithm that adapts sequentially to bandit tasks that it interacts with. The key idea in AdaTS is to adapt to an unknown task prior distribution by maintaining a distribution over its parameters. When solving a bandit task, that uncertainty is marginalized out and properly accounted for. AdaTS is a fully-Bayesian algorithm that can be implemented efficiently in several classes of bandit problems. We derive upper bounds on its Bayes regret that quantify the loss due to not knowing the task prior, and show that it is small. Our theory is supported by experiments, where AdaTS outperforms prior algorithms and works well even in challenging real-world problems.

We propose a bandit algorithm that explores purely by randomizing its past observations. In particular, the sufficient optimism in the mean reward estimates is achieved by exploiting the variance in the past observed rewards. We name the algorithm Capitalizing On Rewards (CORe). The algorithm is general and can be easily applied to different bandit settings. The main benefit of CORe is that its exploration is fully data-dependent. It does not rely on any external noise and adapts to different problems without parameter tuning. We derive a $\tilde O(d\sqrt{n\log K})$ gap-free bound on the n-round regret of CORe in a stochastic linear bandit, where d is the number of features and K is the number of arms. Extensive empirical evaluation on multiple synthetic and real-world problems demonstrates the effectiveness of CORe.

We study two randomized algorithms for generalized linear bandits. The first, GLM-TSL, samples a generalized linear model (GLM) from the Laplace approximation to the posterior distribution. The second, GLM-FPL, fits a GLM to a randomly perturbed history of past rewards. We analyze both algorithms and derive $\tilde{O}(d \sqrt{n \log K})$ upper bounds on their $n$-round regret, where $d$ is the number of features and $K$ is the number of arms. The former improves on prior work while the latter is the first for Gaussian noise perturbations in non-linear models. We empirically evaluate both GLM-TSL and GLM-FPL in logistic bandits, and apply GLM-FPL to neural network bandits. Our work showcases the role of randomization, beyond posterior sampling, in exploration.

We propose RandUCB, a bandit strategy that uses theoretically derived confidence intervals similar to upper confidence bound (UCB) algorithms, but akin to Thompson sampling (TS), uses randomization to trade off exploration and exploitation. In the $K$-armed bandit setting, we show that there are infinitely many variants of RandUCB, all of which achieve the minimax-optimal $\widetilde{O}(\sqrt{K T})$ regret after $T$ rounds. Moreover, in a specific multi-armed bandit setting, we show that both UCB and TS can be recovered as special cases of RandUCB. For structured bandits, where each arm is associated with a $d$-dimensional feature vector and rewards are distributed according to a linear or generalized linear model, we prove that RandUCB achieves the minimax-optimal $\widetilde{O}(d \sqrt{T})$ regret even in the case of infinite arms. We demonstrate the practical effectiveness of RandUCB with experiments in both multi-armed and structured bandit settings. We show that RandUCB matches the empirical performance of TS while matching the theoretically optimal bounds of UCB algorithms, thus achieving the best of both worlds.

We propose a novel framework for structured bandits, which we call an influence diagram bandit. Our framework uses a graphical model to capture complex statistical dependencies between actions, latent variables, and observations; and thus unifies and extends many existing models, such as combinatorial semi-bandits, cascading bandits, and low-rank bandits. We develop novel online learning algorithms that learn to act efficiently in our models. The key idea is to track a structured posterior distribution of model parameters, either exactly or approximately. To act, we sample model parameters from their posterior and then use the structure of the influence diagram to find the most optimistic action under the sampled parameters. We empirically evaluate our algorithms in three structured bandit problems, and show that they perform as well as or better than problem-specific state-of-the-art baselines.

Exploration policies in Bayesian bandits maximize the average reward over problem instances drawn from some distribution P. In this work, we learn such policies for an unknown distribution P using samples from P. Our approach is a form of meta-learning and exploits properties of P without making strong assumptions about its form. To do this, we parameterize our policies in a differentiable way and optimize them by policy gradients, an approach that is pleasantly general and easy to implement. We derive effective gradient estimators and propose novel variance reduction techniques. We also analyze and experiment with various bandit policy classes, including neural networks and a novel softmax policy. The latter has regret guarantees and is a natural starting point for our optimization. Our experiments show the versatility of our approach. We also observe that neural network policies can learn implicit biases expressed only through the sampled instances.

A latent bandit is a bandit problem where the learning agent knows reward distributions of arms conditioned on an unknown discrete latent state. The goal of the agent is to identify the latent state, after which it can act optimally. This setting is a natural midpoint between online and offline learning, where complex models can be learned offline and the agent identifies the latent state online. This is of high practical relevance, for instance in recommender systems. In this work, we propose general algorithms for latent bandits, based on both upper confidence bounds and Thompson sampling. The algorithms are contextual, and aware of model uncertainty and misspecification. We provide a unified theoretical analysis of our algorithms, which have lower regret than classic bandit policies when the number of latent states is smaller than actions. A comprehensive empirical study showcases the advantages of our approach.

Multi-armed bandit (MAB) is a class of online learning problems where a learning agent aims to maximize its expected cumulative reward while repeatedly selecting to pull arms with unknown reward distributions. We consider a scenario where the reward distributions may change in a piecewise-stationary fashion at unknown time steps. We show that by incorporating a simple change-detection component with classic UCB algorithms to detect and adapt to changes, our so-called M-UCB algorithm can achieve nearly optimal regret bound on the order of $O(\sqrt{MKT\log T})$, where $T$ is the number of time steps, $K$ is the number of arms, and $M$ is the number of stationary segments. Comparison with the best available lower bound shows that our M-UCB is nearly optimal in $T$ up to a logarithmic factor. We also compare M-UCB with the state-of-the-art algorithms in numerical experiments using a public Yahoo! dataset and a real-world digital marketing dataset to demonstrate its superior performance.

In many practical problems, a learning agent may want to learn the best action in hindsight without ever taking a bad action, which is much worse than a default production action. In general, this is impossible because the agent has to explore unknown actions, some of which can be bad, to learn better actions. However, when the actions are structured, this is possible if the unknown action can be evaluated by interleaving it with the default action. We formalize this concept as learning in stochastic combinatorial semi-bandits with exchangeable actions. We design efficient learning algorithms for this problem, bound their n-step regret, and evaluate them on both synthetic and real-world problems. Our real-world experiments show that our algorithms can learn to recommend K most attractive movies without ever making disastrous recommendations, both overall and subject to a diversity constraint.

We study sample complexity of optimizing “hill-climbing friendly” functions defined on a graph under noisy observations. We define a notion of convexity, and we show that a variant of best-arm identification can find a near-optimal solution after a small number of queries that is independent of the size of the graph. For functions that have local minima and are nearly convex, we show a sample complexity for the classical simulated annealing under noisy observations. We show effectiveness of the greedy algorithm with restarts and the simulated annealing on problems of graph-based nearest neighbor classification as well as a web advertising application.

We propose a bandit algorithm that explores by randomizing its history of rewards. Specifically, it pulls the arm with the highest mean reward in a non-parametric bootstrap sample of its history with pseudo rewards. We design the pseudo rewards such that the bootstrap mean is optimistic with a sufficiently high probability. We call our algorithm Giro, which stands for garbage in, reward out. We analyze Giro in a Bernoulli bandit and derive a $O(K \Delta^{-1} \log n)$ bound on its $n$-round regret, where $\Delta$ is the difference in the expected rewards of the optimal and the best suboptimal arms, and $K$ is the number of arms. The main advantage of our exploration design is that it easily generalizes to structured problems. To show this, we propose contextual Giro with an arbitrary reward generalization model. We evaluate Giro and its contextual variant on multiple synthetic and real-world problems, and observe that it performs well.

We propose an online algorithm for cumulative regret minimization in a stochastic multi-armed bandit. The algorithm adds O(t) i.i.d. pseudo-rewards to its history in round t and then pulls the arm with the highest average reward in its perturbed history. Therefore, we call it perturbed-history exploration (PHE). The pseudo-rewards are carefully designed to offset potentially underestimated mean rewards of arms with a high probability. We derive near-optimal gap-dependent and gap-free bounds on the n-round regret of PHE. The key step in our analysis is a novel argument that shows that randomized Bernoulli rewards lead to optimism. Finally, we empirically evaluate PHE and show that it is competitive with state-of-the-art baselines.

The rapid growth of mobile devices has resulted in the generation of a large number of user behavior logs that contain latent intentions and user interests. However, exploiting such data in real-world applications is still difficult for service providers due to the complexities of user behavior over a sheer number of possible actions that can vary according to time. In this work, a time-aware RNN model, TRNN, is proposed for predictive analysis from user behavior data. First, our approach predicts next user action more accurately than the baselines including the n-gram models as well as two recently introduced time-aware RNN approaches. Second, we use TRNN to learn user embeddings from sequences of user actions and show that overall the TRNN embeddings outperform conventional RNN embeddings. Similar to how word embeddings benefit a wide range of task in natural language processing, the learned user embeddings are general and could be used in a variety of tasks in the digital marketing area. This claim is supported empirically by evaluating their utility in user conversion prediction, and preferred application prediction. According to the evaluation results, TRNN embeddings perform better than the baselines including Bag of Words (BoW), TFIDF and Doc2Vec. We believe that TRNN embeddings provide an effective representation for solving practical tasks such as recommendation, user segmentation and predictive analysis of business metrics.

Many web systems rank and present a list of items to users, from recommender systems to search and advertising. An important problem in practice is to evaluate new ranking policies offline and optimize them before they are deployed. We address this problem by proposing evaluation algorithms for estimating the expected number of clicks on ranked lists from historical logged data. The existing algorithms are not guaranteed to be statistically efficient in our problem because the number of recommended lists can grow exponentially with their length. To overcome this challenge, we use models of user interaction with the list of items, the so-called click models, to construct estimators that learn statistically efficiently. We analyze our estimators and prove that they are more efficient than the estimators that do not use the structure of the click model, under the assumption that the click model holds. We evaluate our estimators in a series of experiments on a real-world dataset and show that they consistently outperform prior estimators.

Online learning to rank is a sequential decision-making problem where in each round the learning agent chooses a list of items and receives feedback in the form of clicks from the user. Many sample-efficient algorithms have been proposed for this problem that assume a specific click model connecting rankings and user behavior. We propose a generalized click model that encompasses many existing models, including the position-based and cascade models. Our generalization motivates a novel online learning algorithm based on topological sort, which we call TopRank. TopRank is (a) more natural than existing algorithms, (b) has stronger regret guarantees than existing algorithms with comparable generality, (c) has a more insightful proof that leaves the door open to many generalizations, (d) outperforms existing algorithms empirically.

The low-rank structure is one of the most prominent features in modern recommendation problems. In this paper, we consider an online learning problem with a low-rank expected reward matrix where both row features and column features are unknown a priori, and the agent aims to learn to choose the best row-column pair (i.e. the maximum entry) in the matrix. We develop a novel online recommendation algorithm based on ensemble sampling, a recently developed computationally efficient approximation of Thompson sampling. Our computational results show that our algorithm consistently achieves order-of-magnitude improvements over the baselines in both synthetic and real-world experiments.

Modern data streams typically have high dimensionality. For example, digital analytics streams consist of user online activities (e.g., web browsing activity, commercial site activity, apps and social behavior, and response to ads). An important problem is to find frequent joint values (heavy hitters) of subsets of dimensions.

We propose stochastic rank-1 bandits, a class of online learning problems where at each step a learning agent chooses a pair of row and column arms, and receives the product of their values as a reward. The main challenge of the problem is that the individual values of the row and column are unobserved. We assume that these values are stochastic and drawn independently. We propose a computationally-efficient algorithm for solving our problem, which we call Rank1Elim. We derive a O((K + L) (1 / Delta) log n) upper bound on its n-step regret, where K is the number of rows, L is the number of columns, and Delta is the minimum of the row and column gaps; under the assumption that the mean row and column rewards are bounded away from zero. To the best of our knowledge, we present the first bandit algorithm that finds the maximum entry of a rank-1 matrix whose regret is linear in K + L, 1 / Delta, and log n. We also derive a nearly matching lower bound. Finally, we evaluate Rank1Elim empirically on multiple problems. We observe that it leverages the structure of our problems and can learn near-optimal solutions even if our modeling assumptions are mildly violated.

We consider influence maximization (IM) in social networks, which is the problem of maximizing the number of users that become aware of a product by selecting a set of “seed” users to expose the product to. While prior work assumes a known model of information diffusion, we propose a novel parametrization that not only makes our framework agnostic to the underlying diffusion model, but also statistically efficient to learn from data. We give a corresponding monotone, submodular surrogate function, and show that it is a good approximation to the original IM objective. We also consider the case of a new marketer looking to exploit an existing social network, while simultaneously learning the factors governing information propagation. For this, we propose a pairwise-influence semi-bandit feedback model and develop a LinUCB-based bandit algorithm. Our model-independent analysis shows that our regret bound has a better (as compared to previous work) dependence on the size of the network. Experimental evaluation suggests that our framework is robust to the underlying diffusion model and can efficiently learn a near-optimal solution.

Online learning to rank is a core problem in information retrieval and machine learning. Many provably efficient algorithms have been recently proposed for this problem in specific click models. The click model is a model of how the user interacts with a list of documents. Though these results are significant, their impact on practice is limited, because all proposed algorithms are designed for specific click models and lack convergence guarantees in other models. In this work, we propose BatchRank, the first online learning to rank algorithm for a broad class of click models. The class encompasses two most fundamental click models, the cascade and position-based models. We derive a gap-dependent upper bound on the T-step regret of BatchRank and evaluate it on a range of web search queries. We observe that BatchRank outperforms ranked bandits and is more robust than CascadeKL-UCB, an existing algorithm for the cascade model.

The probability that a user will click a search result depends both on its relevance and its position on the results page. The position based model explains this behavior by ascribing to every item an attraction probability, and to every position an examination probability. To be clicked, a result must be both attractive and examined. The probabilities of an item-position pair being clicked thus form the entries of a rank-1 matrix. We propose the learning problem of a Bernoulli rank-1 bandit where at each step, the learning agent chooses a pair of row and column arms, and receives the product of their Bernoulli-distributed values as a reward. This is a special case of the stochastic rank-1 bandit problem considered in recent work that proposed an elimination based algorithm Rank1Elim, and showed that Rank1Elim's regret scales linearly with the number of rows and columns on "benign" instances. These are the instances where the minimum of the average row and column rewards mu is bounded away from zero. The issue with Rank1Elim is that it fails to be competitive with straightforward bandit strategies as mu tends to 0. In this paper we propose Rank1ElimKL, which replaces the crude confidence intervals of Rank1Elim with confidence intervals based on Kullback-Leibler (KL) divergences. With the help of a novel result concerning the scaling of KL divergences we prove that with this change, our algorithm will be competitive no matter the value of mu. Experiments with synthetic data confirm that on benign instances the performance of Rank1ElimKL is significantly better than that of even Rank1Elim. Similarly, experiments with models derived from real-data confirm that the improvements are significant across the board, regardless of whether the data is benign or not.

We study the online influence maximization problem in social networks under the independent cascade model. Specifically, we aim to learn the set of "best influencers" in a social network online while repeatedly interacting with it. We address the challenges of (i) combinatorial action space, since the number of feasible influencer sets grows exponentially with the maximum number of influencers, and (ii) limited feedback, since only the influenced portion of the network is observed. Under a stochastic semi-bandit feedback, we propose and analyze IMLinUCB, a computationally efficient UCB-based algorithm. Our bounds on the cumulative regret are polynomial in all quantities of interest, achieve near-optimal dependence on the number of interactions and reflect the topology of the network and the activation probabilities of its edges, thereby giving insights on the problem complexity. To the best of our knowledge, these are the first such results. Our experiments show that in several representative graph topologies, the regret of IMLinUCB scales as suggested by our upper bounds. IMLinUCB permits linear generalization and thus is both statistically and computationally suitable for large-scale problems. Our experiments also show that IMLinUCB with linear generalization can lead to low regret in real-world online influence maximization.

Weather affects our mood and behavior, and through them, many aspects of our life. When it is sunny, people become happier and smile, but when it rains, some get depressed. Despite this evidence and the abundance of weather data, weather has mostly been overlooked in the machine learning and data science research. This work shows how causal analysis can be applied to discover the effects of weather on TV watching patterns and how it can be applied for user modeling. We make several contributions. First, we show that some weather attributes, e.g., pressure and precipitation, cause significant changes in TV watching patterns. Second, we compare the results obtained for different levels of user granularity and different types of users. This showcases that causal analysis can be a valuable tool in user modeling. To the best of our knowledge, this is the first large-scale causal study of the impact of weather on TV watching patterns.

A search engine recommends to the user a list of web pages. The user examines this list, from the first page to the last, and clicks on all attractive pages until the user is satisfied. This behavior of the user can be described by the dependent click model (DCM). We propose DCM bandits, an online learning variant of the DCM where the goal is to maximize the probability of recommending satisfactory items, such as web pages. The main challenge of our learning problem is that we do not observe which attractive item is satisfactory. We propose a computationally-efficient learning algorithm for solving our problem, dcmKL-UCB; derive gap-dependent upper bounds on its regret under reasonable assumptions; and also prove a matching lower bound up to logarithmic factors. We evaluate our algorithm on synthetic and real-world problems, and show that it performs well even when our model is misspecified. This work presents the first practical and regret-optimal online algorithm for learning to rank with multiple clicks in a cascade-like click model.

Many prior works in recommender systems focus on improving the accuracy of item rating predictions. In comparison, the areas of recommendation interfaces and user-recommender interaction remain underexplored. In this work, we look into the interaction of users with the recommendation list, aiming to devise a method that simplifies content discovery and minimizes the cost of reaching an item of interest. We quantify this cost by the number of user interactions (clicks and scrolls) with the recommendation list. To this end, we propose generalized linear search (GLS), an adaptive combination of the established linear and generalized search (GS) approaches. GLS leverages the advantages of these two approaches, and we prove formally that it performs at least as well as GS. We also conduct a thorough experimental evaluation of GLS and compare it to several baselines and heuristic approaches in both an offline and live evaluation. The results of the evaluation show that GLS consistently outperforms the baseline approaches and is also preferred by users. In summary, GLS offers an efficient and easy-to-use means for content discovery in recommender systems.

A stochastic combinatorial semi-bandit is an online learning problem where at each step a learning agent chooses a subset of ground items subject to constraints, and then observes stochastic weights of these items and receives their sum as a payoff. In this paper, we close the problem of computationally and sample efficient learning in stochastic combinatorial semi-bandits. In particular, we analyze a UCB-like algorithm for solving the problem, which is known to be computationally efficient; and prove O(K L (1 / ∆) \log n) and O(\sqrtK L n \log n) upper bounds on its n-step regret, where L is the number of ground items, K is the maximum number of chosen items, and ∆is the gap between the expected returns of the optimal and best suboptimal solutions. The gap-dependent bound is tight up to a constant factor and the gap-free bound is tight up to a polylogarithmic factor.

A search engine usually outputs a list of K web pages. The user examines this list, from the first web page to the last, and chooses the first attractive page. This model of user behavior is known as the cascade model. In this paper, we propose cascading bandits, a learning variant of the cascade model where the objective is to identify K most attractive items. We formulate our problem as a stochastic combinatorial partial monitoring problem. We propose two algorithms for solving it, CascadeUCB1 and CascadeKL-UCB. We also prove gap-dependent upper bounds on the regret of these algorithms and derive a lower bound on the regret in cascading bandits. The lower bound matches the upper bound of CascadeKL-UCB up to a logarithmic factor. We experiment with our algorithms on several problems. The algorithms perform surprisingly well even when our modeling assumptions are violated.

A stochastic combinatorial semi-bandit is an online learning problem where at each step a learning agent chooses a subset of ground items subject to combinatorial constraints, and then observes stochastic weights of these items and receives their sum as a payoff. In this paper, we consider efficient learning in large-scale combinatorial semi-bandits with linear generalization, and as a solution, propose two learning algorithms called Combinatorial Linear Thompson Sampling (CombLinTS) and Combinatorial Linear UCB (CombLinUCB). Both algorithms are computationally efficient as long as the offline version of the combinatorial problem can be solved efficiently. We establish that CombLinTS and CombLinUCB are also provably statistically efficient under reasonable assumptions, by developing regret bounds that are independent of the problem scale (number of items) and sublinear in time. We also evaluate CombLinTS on a variety of problems with thousands of items. Our experiment results demonstrate that CombLinTS is scalable, robust to the choice of algorithm parameters, and significantly outperforms the best of our baselines.

While numerous works study algorithms for predicting item ratings in recommender systems, the area of the user-recommender interaction remains largely under-explored. In this work, we look into user interaction with the recommendation list, aiming to devise a method that allows users to discover items of interest in a minimal number of interactions. We propose generalized linear search (GLS), a combination of linear and generalized searches that brings together the benefits of both approaches. We prove that GLS performs at least as well as generalized search and compare our method to several baselines and heuristics. Our evaluation shows that GLS is liked by the users and achieves the shortest interactions.

Matrix factorization (MF) collaborative filtering is an effective and widely used method in recommendation systems. However, the problem of finding an optimal trade-off between exploration and exploitation (otherwise known as the bandit problem), a crucial problem in collaborative filtering from cold-start, has not been previously addressed.In this paper, we present a novel algorithm for online MF recommendation that automatically combines finding the most relevantitems with exploring new or less-recommended items.Our approach, called Particle Thompson Sampling for Matrix-Factorization, is based on the general Thompson sampling framework, but augmented with a novel efficient online Bayesian probabilistic matrix factorization method based on the Rao-Blackwellized particle filter.Extensive experiments in collaborative filtering using several real-world datasets demonstrate that our proposed algorithm significantly outperforms the current state-of-the-arts.

We propose combinatorial cascading bandits, a class of partial monitoring problems where at each step a learning agent chooses a tuple of ground items subject to constraints and receives a reward if and only if the weights of all chosen items are one. The weights of the items are binary, stochastic, and drawn independently of each other. The agent observes the index of the first chosen item whose weight is zero. This observation model arises in network routing, for instance, where the learning agent may only observe the first link in the routing path which is down, and blocks the path. We propose a UCB-like algorithm for solving our problems, CombCascade; and prove gap-dependent and gap-free upper bounds on its n-step regret. Our proofs build on recent work in stochastic combinatorial semi-bandits but also address two novel challenges of our setting, a non-linear reward function and partial observability. We evaluate CombCascade on two real-world problems and show that it performs well even when our modeling assumptions are violated. We also demonstrate that our setting requires a new learning algorithm.

Maximization of submodular functions has wide applications in artificial intelligence and machine learning. In this paper, we propose a scalable learning algorithm for maximizing an adaptive submodular function. The key structural assumption in our solution is that the state of each item is distributed according to a generalized linear model, which is conditioned on the feature vector of the item. Our objective is to learn the parameters of this model. We analyze the performance of our algorithm, and show that its regret is polylogarithmic in time and linear in the number of features. Finally, we evaluate our solution on two problems, preference elicitation and adaptive face detection, and demonstrate that high-quality policies can be learned sample efficiently.

Smooth functions on graphs have wide applications in manifold and semi-supervised learning. In this paper, we study a bandit problem where the payoffs of arms are smooth on a graph. This framework is suitable for solving online learning problems that involve graphs, such as content-based recommendation. In this problem, each item we can recommend is a node and its expected rating is similar to its neighbors. The goal is to recommend items that have high expected ratings. We aim for the algorithms where the cumulative regret with respect to the optimal policy would not scale poorly with the number of nodes. In particular, we introduce the notion of an effective dimension, which is small in real-world graphs, and propose two algorithms for solving our problem that scale linearly and sublinearly in this dimension. Our experiments on real-world content recommendation problem show that a good estimator of user preferences for thousands of items can be learned from just tens of nodes evaluations.

Kernel-based reinforcement learning (KBRL) is a popular approach to learning non-parametric value function approximations. In this paper, we present structured KBRL, a paradigm for kernel-based RL that allows for modeling independencies in the transition and reward models of problems. Real-world problems often exhibit this structure and can be solved more efficiently when it is modeled. We make three contributions. First, we motivate our work, define a structured backup operator, and prove that it is a contraction. Second, we show how to evaluate our operator efficiently. Our analysis reveals that the fixed point of the operator is the optimal value function in a special factored MDP. Finally, we evaluate our method on a synthetic problem and compare it to two KBRL baselines. In most experiments, we learn better policies than the baselines from an order of magnitude less training data.

Millions of people search daily for movies, music, and books on the Internet. Unfortunately, non-personalized exploration of items can result in an infeasible number of costly interaction steps. We study the problem of efficient, repeated interactive search. In this problem, the user is navigated to the items of interest through a series of options and our objective is to learn a better search policy from past interactions with the user. We propose an efficient learning algorithm for solving the problem, sequential Bayesian search (SBS), and prove that it is Bayesian optimal. We also analyze the algorithm from the frequentist point of view and show that its regret is sublinear in the number of searches. Finally, we evaluate our method on a real-world movie discovery problem and show that it performs nearly optimally as the number of searches increases.

We design predictors of user dissatisfaction with the performance of applications that use networking. Our approach combines user-level feedback with low level machine and networking metrics. The main challenges of predicting user dissatisfaction, that arises when networking conditions adversely affect applications, comes from the scarcity of user feedback and the fact that poor performance episodes are rare. We develop a methodology to handle these challenges. Our method processes low level data via quantization and feature selection steps. We combine this with user labels and employ supervised learning techniques to build predictors. Using data from 19 personal machines, we show how to build training sets and demonstrate that non-linear SVMs achieve higher true positive rates (around 0.9) than predictors based on linear models. Finally we quantify the benefits of building per-application predictors as compared to general predictors that use data from multiple applications simultaneously to anticipate user dissatisfaction.

Maximization of submodular functions has wide applications in machine learning and artificial intelligence. Adaptive submodular maximization has been traditionally studied under the assumption that the model of the world, the expected gain of choosing an item given previously selected items and their states, is known. In this paper, we study the scenario where the expected gain is initially unknown and it is learned by interacting repeatedly with the optimized function. We propose an efficient algorithm for solving our problem and prove that its expected cumulative regret increases logarithmically with time. Our regret bound captures the inherent property of submodular maximization, earlier mistakes are more costly than later ones. We refer to our approach as Optimistic Adaptive Submodular Maximization (OASM) because it trades off exploration and exploitation based on the optimism in the face of uncertainty principle. We evaluate our method on a preference elicitation problem and show that non-trivial K-step policies can be learned from just a few hundred interactions with the problem.

Markov decision processes (MDPs) are an established framework for solving sequential decision-making problems under uncertainty. In this work, we propose a new method for batch-mode reinforcement learning (RL) with continuous state variables. The method is an approximation to kernel-based RL on a set of k representative states. Similarly to kernel-based RL, our solution is a fixed point of a kernelized Bellman operator and can approximate the optimal solution to an arbitrary level of granularity. Unlike kernel-based RL, our method is fast. In particular, our policies can be computed in O(n) time, where n is the number of training examples. The time complexity of kernel-based RL is Ω(n2). We introduce our method, analyze its convergence, and compare it to existing work. The method is evaluated on two existing control problems with 2 to 4 continuous variables and a new problem with 64 variables. In all cases, we outperform state-of-the-art results and offer simpler solutions.

In this paper, we consider the problem of conditional anomaly detection that aims to identify data instances with an unusual response or a class label. We develop a new non-parametric approach for conditional anomaly detection based on the soft harmonic solution, with which we estimate the confidence of the label to detect anomalous mislabeling. We further regularize the solution to avoid the detection of isolated examples and examples on the boundary of the distribution support. We demonstrate the efficacy of the proposed method on several synthetic and UCI ML datasets in detecting unusual labels when compared to several baseline approaches. We also evaluate the performance of our method on a real-world electronic health record dataset where we seek to identify unusual patient-management decisions.

Non-invasive identity inference in the home environment is a very challenging problem. A practical solution to the problem could have far reaching implications in many industries, such as home entertainment. In this work, we consider the problem of identity inference using a TV remote control. In particular, we address two challenges that have so far prevented the work of Chang et al. (2009) from being applied in a home entertainment system. First, we show how to learn the patterns of TV remote controls incrementally and online. Second, we generalize our results to partially labeled data. To achieve our goal, we use state-of-the-art methods for max-margin learning and online convex programming. Our solution is efficient, runs in real time, and comes with theoretical guarantees. It performs well in practice and we demonstrate this on 4 datasets of 2 to 4 people.

Efficient representations and solutions for large decision problems with continuous and discrete variables are among the most important challenges faced by the designers of automated decision support systems. In this paper, we describe a novel hybrid factored Markov decision process (MDP) model that allows for a compact representation of these problems, and a new hybrid approximate linear programming (HALP) framework that permits their efficient solutions. The central idea of HALP is to approximate the optimal value function by a linear combination of basis functions and optimize its weights by linear programming. We analyze both theoretical and computational aspects of this approach, and demonstrate its scale-up potential on several hybrid optimization problems.

Approximate linear programming (ALP) has emerged recently as one of the most promising methods for solving complex factored MDPs with (cid:2)nite state spaces. In this work we show that ALP solutions are not limited only to MDPs with (cid:2)nite state spaces, but that they can also be applied successfully to factored continuous-state MDPs (CMDPs). We show how one can build an ALP-based approximation for such a model and contrast it to existing solution methods. We argue that this approach offers a robust alternative for solving high dimensional continuous-state space problems. The point is supported by experiments on three CMDP problems with 24-25 continuous state factors.