New approach for distributed online optimization of non-convex losses with sublinear regret.
problem Regret evaluation and consensus in distributed, multi-agent systems with non-convex losses.
method Composite regret metric and consensus-based online normalized gradient (CONGD) approach for pseudo-convex losses; offline optimization oracle for general non-convex losses.
result First sublinear regret bound for general distributed online non-convex learning.
Proposes MRO to achieve uniformly low regret in distributionally robust learning.
problem Learning under unknown test distributions (distribution shift).
method Minimax Regret Optimization (MRO) for robust machine learning.
result MRO achieves uniformly low regret across all test distributions.
Unified framework for distributional regret in bandits and reinforcement learning.
problem Characterizing the distribution of regret in multi-armed bandits and reinforcement learning.
method Unified framework with a UCBVI-style algorithm and distributional regret bounds.
result Distributional regret bounds with optimal trade-offs between expected and distributional regret.
Bandit algorithms struggle with consistent performance and robustness.
problem Achieving consistent and robust performance in stochastic multi-armed bandit settings.
method Analyzing regret minimization trade-offs and proposing distribution-oblivious algorithms.
result Logarithmic regret is inconsistent and super-logarithmic regret is necessary for consistent learning.
New algorithm SELECT minimizes satisficing regret in bandits.
problem Minimizing regret in bandit optimization with satisficing arms.
method SELECT algorithm for satisficing regret minimization.
result SELECT achieves constant expected satisficing regret.
New algorithm reduces individual regret and communication costs in cooperative bandits.
problem Optimal individual and group regret in cooperative multi-agent bandits.
method Integrates a new communication policy into a learning algorithm.
result Achieves optimal individual regret and constant communication costs.
New algorithm reduces worst-case regret for heavy-tailed bandits.
problem Stochastic Multi-Armed Bandit problem with heavy-tailed rewards.
method Modified minimax policy MOSS with saturated empirical mean.
result Worst-case regret matching lower bound for heavy-tailed distributions.
This paper analyzes data-driven Newsvendor problems and finds a wide range of possible regrets.
problem Guessing the number drawn from an unknown distribution with asymmetric costs.
method Unified analysis using the notion of clustered distributions and new lower bounds.
result The entire spectrum of achievable regrets from 1 / n 1/\sqrt{n} 1/ n to 1 / n 1/n 1/ n is possible. Greedy algorithm achieves sublinear regret for various distributions.
problem Efficient performance of greedy algorithms in linear contextual bandit problems.
method Introduced Local Anti-Concentration (LAC) condition to ensure sublinear regret.
result Greedy algorithm achieves O ( poly log T ) O(\operatorname{poly} \log T) O ( poly log T ) cumulative expected regret. New trade-off found in bandit problems with unknown range.
problem Stochastic bandit problems with unknown range.
method Exhibit a strategy achieving new trade-off between distribution-dependent and distribution-free regret bounds.
result Achieves the rates for regret indicated by the new trade-off.
We study the problem of regret minimization for distributed bandits learning, in which M M M agents work collaboratively to minimize their total regret under the coordination of a central server. Our goal is to design communication protocols with near-optimal regret and little communication cost, which is measured by the…
We consider K K K -armed stochastic bandits and consider cumulative regret bounds up to time T T T . We are interested in strategies achieving simultaneously a distribution-free regret bound of optimal order K T \sqrt{KT} K T and a distribution-dependent regret that is asymptotically optimal, that is, matching the κ ln T κ\ln T κ ln T lower b…
New Thompson sampling algorithm reduces regret for exponential family bandits.
problem Minimizing regret in multi-armed bandit problems with exponential family rewards.
method Proposes ExpTS and ExpTS + ^+ + algorithms using novel sampling distributions. result Minimizes both finite-time and asymptotic regret for exponential family rewards.
LaPSRL achieves optimal regret for isoperimetric RL distributions.
problem Designing RL algorithms with sublinear regret for non-log-concave distributions.
method Posterior Sampling (PSRL) and Langevin sampling (LaPSRL) for isoperimetric distributions.
result LaPSRL achieves order-optimal regret and subquadratic complexity.
Optimal learning rate schedules for SGD in changing data distributions.
problem Minimizing regret in online learning with changing data distributions.
method Characterized optimal schedules for linear regression, proposed schedules for general convex and non-convex losses, and defined a notion of regret for non-convex losses.
result Upper and lower bounds for regret with constants for convex losses, and an upper bound on total expected regret for non-convex losses.
Proposes a general method to derive regret bounds for multi-armed bandit algorithms.
problem Deriving regret bounds for randomized multi-armed bandit algorithms.
method Checking sufficient conditions on sampling probabilities and distributions.
result Proves logarithmic regret bounds for various bandit algorithms and new models.
CRIMED optimizes regret in bandits with unbounded stochastic corruption.
problem Minimizing regret in bandits with arbitrary unbounded corruptions.
method Introduces CRIMED, an asymptotically-optimal algorithm for Gaussian distributions with known variance.
result Achieves exact lower bound on regret for Gaussian distributions with high corruption probability.
Optimized bandit algorithms have heavy-tailed regret distributions that can grow faster than expected.
problem Heavy-tailed regret distributions in optimized bandit algorithms.
method Change-of-measure ideas and UCB algorithm modifications.
result Regret distributions of optimized UCB algorithms have a heavy Cauchy tail, and can grow faster than poly-logarithmically.
A new policy for contextual bandits adapts to reward vector shifts.
problem Learning under reward vector shifts with ordered rewards.
method Adaptive-discretization and optimistic elimination policy.
result Established upper bounds on preference-based regret.
Optimal algorithm for minimizing regret in heavy-tailed bandits.
problem Minimizing regret in stochastic multi-armed bandits with heavy-tailed distributions.
method Proposes an optimal algorithm under the assumption of uniformly bounded moments of order (1+ε).
result Matches the lower bound exactly in the first-order term and provides a finite-time bound on its regret.
New meta-learning framework for minimizing simple regret in bandits.
problem Minimizing simple regret in a sequence of bandit tasks with unknown distributions.
method Developed Bayesian and frequentist meta-learning algorithms for bandits, analyzing their meta simple regret.
result Bayesian algorithm achieves i l d e O ( m / n ) ilde{O}(m / \sqrt{n}) i l d e O ( m / n ) meta simple regret, frequentist algorithm i l d e O ( m n + m / n ) ilde{O}(\sqrt{m} n + m/ \sqrt{n}) i l d e O ( m n + m / n ) . Optimal algorithm for identifying best-arm with minimal regret.
problem Identifying the best arm in two treatments with limited budget.
method Neyman allocation based on outcome standard deviations.
result Neyman allocation is minimax optimal for simple regret.
Learning reward functions can lead to poor policy performance despite low error.
problem Low error in learned reward functions does not guarantee low regret in policy performance.
method Mathematical analysis of reward learning and policy optimization.
result A low expected test error of the reward model guarantees low worst-case regret, but error-regret mismatch can occur with certain data distributions.
The problem of distributed learning and channel access is considered in a cognitive network with multiple secondary users. The availability statistics of the channels are initially unknown to the secondary users and are estimated using sensing decisions. There is no explicit information exchange or prior agreement amon…
Study on sequential prediction with log-loss, focusing on well-specified and misspecified cases.
problem Sequential prediction with log-loss under different specification conditions.
method Analysis of cumulative regret in well-specified and misspecified cases for a Gaussian location hypothesis class.
result Cumulative regrets in well-specified and misspecified cases asymptotically coincide for the d d d -dimensional Gaussian location hypothesis class. Algorithm minimizes regret in predictive models influenced by their own predictions.
problem Finding near-optimal models under performativity with unknown shifts.
method Developed an algorithm that uses performative feedback to achieve low regret, scaling only with distribution shift complexity.
result Achieved regret bounds scaling with distribution shift complexity, not reward function complexity.
Most bandit algorithm designs are purely theoretical. Therefore, they have strong regret guarantees, but also are often too conservative in practice. In this work, we pioneer the idea of algorithm design by minimizing the empirical Bayes regret, the average regret over problem instances sampled from a known distributio…
New algorithms reduce risk in reinforcement learning with provable regret bounds.
problem Risk-sensitive reinforcement learning in Markov decision processes.
method Two novel DRL algorithms leveraging the independence property of entropic risk measure.
result Regret bounds of i l d e O ( exp ( ∣ β ∣ H ) − 1 ∣ β ∣ H S 2 A K ) ilde{\mathcal{O}}(\frac{\exp(|β| H)-1}{|β|}H\sqrt{S^2AK}) i l d e O ( ∣ β ∣ e x p ( ∣ β ∣ H ) − 1 H S 2 A K ) for model-free and model-based algorithms. A method to minimize regret in multi-agent control systems with adversarial disturbances.
problem Optimal control of dynamical systems with adversarial disturbances and multiple agents.
method Reduction from online convex optimization to a distributed algorithm for multi-agent control.
result The resulting distributed algorithm has low regret relative to the optimal precomputed joint policy.
Recent work on follow the perturbed leader (FTPL) algorithms for the adversarial multi-armed bandit problem has highlighted the role of the hazard rate of the distribution generating the perturbations. Assuming that the hazard rate is bounded, it is possible to provide regret analyses for a variety of FTPL algorithms f…
We consider the stochastic multi-armed bandit problem with a prior distribution on the reward distributions. We are interested in studying prior-free and prior-dependent regret bounds, very much in the same spirit as the usual distribution-free and distribution-dependent bounds for the non-Bayesian stochastic bandit. B…
Paper improves regret bounds for distributed experts problem.
problem Minimizing loss in a distributed experts problem.
method Protocol achieving improved regret bound with minimized communication.
result Regret bound improved to R ≳ 1 T ⋅ e x t p o l y log ( n s T ) R \gtrsim \frac{1}{\sqrt{T} \cdot ext{poly}\log(nsT)} R ≳ T ⋅ e x t p o l y l o g ( n s T ) 1 . New online conformal prediction methods minimize strongly adaptive regret and achieve near-optimal coverage.
problem Uncertainty quantification in online settings with changing data distributions.
method Developed new online conformal prediction methods that minimize strongly adaptive regret.
result Achieve near-optimal strongly adaptive regret and approximately valid coverage.
Algorithm reduces regret in distributed kernel bandits with shared randomness.
problem Minimizing regret in collaborative function maximization.
method Uniform exploration at local agents and shared randomness with central server.
result Achieves optimal regret order with sublinear communication cost.
Distributed Thompson sampling improves regret convergence in constrained communication networks.
problem Maximizing a black-box function with multi-agent Bayesian optimization under communication constraints.
method Distributed Thompson sampling using Gaussian processes, with theoretical bounds on regret convergence.
result Theoretical bounds on Bayesian average and simple regret depend on communication graph structure and are applicable in constrained networks.
This paper improves GP-UCB by using a shifted exponential distribution for confidence parameters.
problem Theoretical confidence parameter in GP-UCB increases with iterations, leading to large values.
method Introduced IRGP-UCB, a randomized variant of GP-UCB using a shifted exponential distribution for confidence parameters.
result IRGP-UCB achieves sub-linear regret without increasing the confidence parameter.
A robust bandit algorithm uses Dirichlet sampling to minimize regret under various distributional assumptions.
problem Robustness of bandit algorithms to model misspecification.
method Dirichlet Sampling (DS) algorithm based on pairwise comparisons and re-sampling of arm observations.
result Different DS variants achieve optimal regret guarantees for bounded distributions and logarithmic regret for semi-bounded distributions.
A novel algorithm minimizes regret in a multi-agent bandit problem with time-varying random graphs and heterogeneous rewards.
problem Minimizing regret in a multi-agent multi-armed bandit problem with time-varying random graphs and heterogeneous rewards.
method Introduces a novel algorithmic framework combining averaging-based consensus with a weighting technique and upper confidence bound.
result Derives optimal instance-dependent regret upper bounds of order log T \log{T} log T in both sub-gaussian and sub-exponential environments. In this paper, we introduce Ballooning Multi-Armed Bandits (BL-MAB), a novel extension of the classical stochastic MAB model. In the BL-MAB model, the set of available arms grows (or balloons) over time. In contrast to the classical MAB setting where the regret is computed with respect to the best arm overall, the regr…
New algorithm DUSA minimizes regret in structured bandits with structural information.
problem Optimal decision-making under uncertainty with structural reward information.
method DUSA algorithm exploiting convex structural information.
result Regret matches information-theoretic lower bound up to a constant factor.
Adaptive algorithms minimize regret in matching markets with contextual arm preferences.
problem Minimizing regret in matching markets with context-dependent player utilities.
method Developed adaptive algorithms for stochastic and adversarial contexts, providing upper and lower bounds.
result Achieved sublinear regret bounds for both stochastic and adversarial contexts.
New algorithm reduces online learning error for unknown feature distributions.
problem Oracle-efficient hybrid online learning with unknown feature and label distributions.
method Computational efficient online predictor using ERM oracle for finite-VC and fat-shattering classes.
result Oracle-efficient sublinear regret bounds for hybrid online learning with unknown feature generation.
Thompson Sampling shows polynomial regret for combinatorial semi-bandits with subgaussian rewards.
problem Finding optimal solutions in combinatorial semi-bandits with suboptimal sampling.
method Proposes Thompson Sampling with polynomial regret for linear combinatorial semi-bandits.
result Demonstrates 'mismatched sampling paradox' where knowing distributions can lead to worse performance.
Optimizes regret distribution in stochastic bandits for risk balance.
problem Balancing regret expectation and tail risk in stochastic bandits.
method Characterizes optimal regret tail probability for any threshold, proposes new policies.
result Discovers an intrinsic gap in optimal tail rate based on time horizon uncertainty.
Improved Thompson Sampling using fractional posteriors achieves better regret bounds.
problem Optimizing regret in stochastic multi-armed bandit problems.
method Using α \alpha α -posterior distributions, derived frequentist regret bounds. result Instance-dependent and instance-independent regret bounds established.
New method for distributed online learning with communication constraints reduces joint regret.
problem Joint regret minimization in a distributed online learning setting with communication constraints.
method Adaptive graph partitioning and comparator-adaptive online convex optimization with delayed gradient information.
result Optimal graph partition selection for adversarial activations and gradients reduces joint regret.
SCaLE tackles dynamic regret in noisy bandit feedback with switching costs.
problem Unbounded metric movement costs in bandit online convex optimization.
method SCaLE algorithm for high-dimensional dynamic quadratic hitting costs and ℓ 2 \ell_2 ℓ 2 -norm switching costs, with spectral regret analysis. result First algorithm achieving sub-linear dynamic regret without hitting cost knowledge.
The stochastic multi-armed bandit problem is well understood when the reward distributions are sub-Gaussian. In this paper we examine the bandit problem under the weaker assumption that the distributions have moments of order 1+ε, for some ε ∈ ( 0 , 1 ] ε\in (0,1] ε ∈ ( 0 , 1 ] . Surprisingly, moments of order 2 (i.e., finite variance) are suffi…