Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 20 of 24 for “"Bandit feedback"”.
-
Adaptive Recommendations with Bandit Feedback
… in machine learning. We consider multi-armed bandits problems, where the goal is to design a policy to recommend arms (options) sequentially where in each round only the rewards sampled from the selected arms are available. In this thesis, we study both theoretical and practical aspects of …
-
Submodular Cover and Maximization: Approximation Algorithms, Bandit Feedback, and Reinforcement Learning
… of submodular optimization under stochastic bandit feedback, where the function is accessed only through noisy queries drawn from an unknown distribution. This dissertation aims to addresses these challenges through contributions across both problem domains. For submodular cover, we present …
-
Advances in online convex optimization, games, and problems with bandit feedback
… we study a problem called: Risk-Averse Convex Bandit. Risk-aversion makes reference to the fact that humans prefer consistent sequences of good rewards instead of highly variable sequences with slightly better rewards. The Risk-Averse Convex Bandit addresses the fact that, while human decision …
-
Learning Posted Prices in Bilateral Trade: Regret Guarantees Under Full and Bandit Feedback
… (e.g., a posted price) and observes structured feedback. We ask how the information revealed after each decision determines whether learning is possible and what regret rates are achievable. We cast the problem in the online-learning framework and analyze two feedback models. Under …
-
Adaptive Preference Learning With Bandit Feedback: Information Filtering, Dueling Bandits and Incentivizing Exploration
… learning system learns users' preferences from feedback while simultaneously using these learned preferences to help them find preferred items. We study three different types of user feedback in three application setting: cardinal feedback with application in information filtering systems, …
-
Bandit learning with non-linear rewards: Submodularity and beyond
… focuses on the Combinatorial Multi-Armed Bandit (CMAB) problem, especially in the context where the reward function has non-linear characteristics and the feedback is bandit feedback. The main contributions include proposing the Explore-Then-Commit Greedy (ETCG) algorithm to solve the CMAB …
-
Multi-armed bandits and applications to large datasets
This thesis considers the multi-armed bandit (MAB) problem, both the traditional bandit feedback and graphical bandits when there is side information. Motivated by the Boltzmann exploration algorithm often used in the more general context of reinforcement learning, we present Almost Boltzmann …
-
Online and active learning of big networks: theory and algorithms
… active learning), and online learning with bandit feedback algorithms for learning in a network. In the first part of this thesis, I propose a \textit{nonadaptive} active learning approach on a graph, based on generalization error bound minimization. In particular, I present a data-dependent …
-
Structured bandits and applications : exploiting problem structure for better decision-making under uncertainty
… of decision-making under uncertainty in the bandit setting. This thesis goes beyond the well-studied multi-armed bandit model to consider structured bandit settings and their applications. In particular, we learn to make better decisions by leveraging the application-specific …
-
Swap-regret-minimizing bandits for distributed network optimization
… network optimization grounded in game-theoretic bandit learning. We model fundamental problems, such as congestion control and resource allocation, as repeated games involving strategic agents who receive only partial (bandit) feedback. Motivated by practical challenges in computer networks, we …
-
Pessimistic-Optimistic Bandit Learning with Applications to Communication Networks
… statistics under various constraints with bandit feedback. In this thesis, we study the pessimistic-optimistic approach to solving extensions to constrained bandit learning, motivated by various communication network and other resource allocation applications. In particular, we study the …
-
Algorithms for Online Marketplaces: New Approaches to Order Fulfillment and Recommendation Systems
… recommendation systems through a contextual bandit framework, examining both full-feedback and bandit-feedback settings. By formulating the problem to accommodate arbitrary mappings from user contexts to product feature values, this research provides new online algorithms that effectively …
-
Constrained learning in the bandit setting: doubly optimistic strategies and fast rates
The (stochastic) bandit problem is a classic example used to address the challenge of balancing exploration and exploitation when dealing with bandit feedback. This dissertation focuses on stochastic bandits with constraints, where each action (or arm) bears both a reward and a safety risk. The …
-
COMPETING AGAINST ADAPTIVE AGENTS BY MINIMIZING COUNTERFACTUAL NOTIONS OF REGRET
… This thesis studies policy regret in a partial (bandit) feedback environment, beyond the worst case setting, by leveraging additional structure such as stochasticity/stability of the adversary or additional feedback.
-
Learning with Minimal Supervision: New Meta-Learning and Reinforcement Learning Algorithms
… policies in the interactive contextual bandit setting. All of these algorithms simulate settings in which the agent has access to only a few labeled samples. Based on these simulations, the agent learns how to solve future learning tasks with minimal supervision. In the second part of …
-
Online machine learning framework for budgeted bandits with an option of giving up
… model this problem as a variant of multi-armed bandits problem having stochastic reward and stochastic resource consumption with a fixed global budget. For this problem, we first establish that the optimal arm is the arm that maximizes the ratio of the expected reward of the arm to the expected …
-
Ensembling and Coordination Methods for Sequential Decision Making Under Uncertainty
… this structure. This work addresses the field of bandit modeling, providing methods for ensembling bandit algorithms and for coordinating simultaneous decisions with bandit feedback. We provide both theoretical and empirical justifications for our methods in the form of simulation studies, …
-
Computationally bounded rationality from three perspectives: precomputation, regret tradeoffs, and lifelong learning
… regret guarantee in both the full and semi-bandit feedback settings. We then show how this tradeoff implies new results for Online Submodular Function Maximization (OSFM) (Streeter and Golovin, 2008) and Linear Programming. We use these observations to derive and benchmark a new algorithm …
-
Efficient Robust Algorithms for Linear Discriminant Analysis and Sequential Matching Problems
… proposed algorithm is to use stochastic binary bandit feedback to adaptively estimate the unknown utilities through the logistic regression, and then to combine the Upper Confidence Bound (UCB) algorithm in the multi-armed bandit problem with the Hungarian algorithm in the assignment problem. We …
-
Counterfactual Evaluation And Learning From Logged User Feedback
… as batch learning from logged contextual bandit feedback. We will first develop better off-policy estimators for evaluating online user-centric metrics in information retrieval applications. In subsequent chapters, we will study the bias-variance trade-off when learning from logged …
Page 1 of 2