New policies maximize rewards in social networks with side-observation data.
problem Maximizing rewards in stochastic multi-armed bandit problems with network side-observations.
method Proposed two policies: randomized and UCB-based, achieving asymptotic lower bound on regret.
result Policies achieve maximum long-term average reward up to a multiplicative factor, independent of network structure.
A new algorithm for social network recommendations using side-observations.
problem Designing recommendation algorithms for users influenced by their social network.
method Contextual bandits with side-observations modeled by a social network graph.
result The proposed algorithm achieves asymptotically optimal regret, matching the lower-bound as To∞. Two algorithms minimize regret in adversarial bandit problems with side-observation losses.
problem Minimizing regret in adversarial multi-armed bandit problems with side-observation losses.
method Proposes two algorithms for different ranges of side-observation probability.
result Regret bounds for different values of side-observation probability.
New algorithm reduces regret in bandits with occasional free observations.
problem Reducing regret in bandit problems with occasional free observations.
method Developed an algorithm with a regret bound of Σ_i (log(1/ε) / Δ_i) up to constants and loglog terms.
result Proved that the algorithm's regret is optimal, matching lower bounds.
Study invariant Lipschitz bandits, improving regret bounds.
problem Optimizing decisions under symmetry in online settings.
method Integrates side observations using group orbits into UniformMesh algorithm.
result Improved regret bound for invariant Lipschitz bandit class.
This paper considers stochastic bandits with side observations, a model that accounts for both the exploration/exploitation dilemma and relationships between arms. In this setting, after pulling an arm i, the decision maker also observes the rewards for some other actions related to i. We will see that this model is su…
We consider an adversarial online learning setting where a decision maker can choose an action in every stage of the game. In addition to observing the reward of the chosen action, the decision maker gets side observations on the reward he would have obtained had he chosen some of the other actions. The observation str…
Investigates sequential problems on graph structures and large action spaces.
problem Sequential decision-making on graph structures and large action spaces.
method Spectral bandits, side observations, influence maximization, kernel bandits, polymatroid bandits, function optimization, infinitely many-arms bandits.
result Contributions to graph and structured bandits.
New algorithm for online learning with noisy side observations.
problem Online learning with noisy side feedback and graph-structured dependencies.
method Proposes an algorithm using a weighted directed graph to model dependencies and guarantees a regret bound of O(√α* T).
result Guarantees a regret of O(√α* T) after T rounds, where α* is the effective independence number.
Contextual linear optimization shows naive plug-in methods can outperform direct optimization.
problem Optimizing decisions with side observations to reduce uncertainty.
method Using off-the-shelf machine learning methods to learn a predictive model and plug it in for optimization.
result The naive plug-in approach achieves faster regret convergence rates than direct optimization methods.
We consider a sequential learning problem with Gaussian payoffs and side information: after selecting an action i, the learner receives information about the payoff of every action j in the form of Gaussian observations whose mean is the same as the mean payoff, but the variance depends on the pair (i,j) (and may…
A network of agents attempt to learn some unknown state of the world drawn by nature from a finite set. Agents observe private signals conditioned on the true state, and form beliefs about the unknown state accordingly. Each agent may face an identification problem in the sense that she cannot distinguish the truth in …
Optimizes arm selection with side information in Gaussian bandits.
problem Optimizing arm selection with side information in Gaussian bandits.
method Constructs an LP-based asymptotic instance-dependent lower bound on the regret and develops the first known asymptotically optimal algorithm.
result First known asymptotically optimal algorithm for Gaussian bandits with side information.
Best-arm identification in bandits with sequential elimination.
problem Identifying the arm with the highest expected reward in a budget-limited exploration process.
method Unified sequential elimination algorithms, dividing budget based on nonlinear function of remaining arms.
result Improved theoretical guarantees and performance over state-of-the-art algorithms.
The holographic duality can be extended to include quantum theories with broken coordinate invariance leading to the appearance of the gravitational anomalies. On the gravity side one adds the gravitational Chern-Simons term to the bulk action which gauge invariance is only up to the boundary terms. We analyze in detai…
New algorithms for efficient learning with partial information, reducing regret.
problem Online learning with partial observability and semi-bandit feedback.
method Implicit exploration strategy for near-optimal regret guarantees.
result First algorithms with near-optimal regret guarantees without knowing the observation system.