The paper derives a new theorem for predicting batches of data.
problem Finding lower bounds on minimal batch regret.
method Derives a conditional version of the regret-capacity theorem.
result Reveals a connection between conditional Rényi divergence and conditional Sibson's mutual information.
New insights link no-regret learning to online conformal prediction in adversarial settings.
problem Understanding the relationship between no-regret learning and online conformal prediction in adversarial environments.
method Analysis of existing algorithms and new connections between no-regret learning and conformal prediction.
result No-regret learning algorithms can provide group-conditional coverage guarantees in adversarial settings.
New algorithms achieve logarithmic regret in learning linear quadratic control systems.
problem Learning in Linear Quadratic Control systems with unknown parameters.
method Efficient algorithms for two scenarios: unknown A A A or B B B with certain conditions. result Regret scales logarithmically with the number of steps, not square root.
New algorithm reduces bandit problem's regret bound to logarithmic in dimension.
problem Sparse linear bandit problem with sparse reward structure.
method Proposes an algorithm that uses compatibility condition on optimal arm.
result Achieves regret bound of O(poly log dT) without additional diversity assumptions.
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. Paper improves CMAB regret bounds by reducing batch-size dependency.
problem Reducing batch-size dependency in combinatorial semi-bandits.
method Developed BCUCB-T and SESCB algorithms with new TPVM conditions.
result Significantly improved regret bounds for various applications.
Enhanced H H H -consistency bounds derived under relaxed conditions.
problem Quantifying the relationship between zero-one estimation error and surrogate loss estimation error.
method Relaxing the condition on the surrogate loss conditional regret and presenting a general framework for establishing enhanced H H H -consistency bounds. result Derivation of more favorable H H H -consistency bounds in various scenarios. 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…
New coverage conditions improve sample efficiency in online reinforcement learning.
problem Improving sample efficiency in online reinforcement learning with function approximation.
method Identifying and studying new coverage conditions for online reinforcement learning.
result Improved regret bounds achieved with new coverage conditions.
The paper analyzes the sliding regret of stochastic bandit algorithms.
problem Measuring the one-shot behavior of no-regret algorithms in stochastic bandits.
method Introducing sliding regret to measure the worst pseudo-regret over a time-window.
result Randomized methods have optimal sliding regret, while index policies have the worst possible sliding regret.
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.
We investigate online convex optimization in changing environments, and choose the adaptive regret as the performance measure. The goal is to achieve a small regret over every interval so that the comparator is allowed to change over time. Different from previous works that only utilize the convexity condition, this pa…
Study online conformal prediction for non-stationary data with optimal training-conditional regret.
problem Online prediction for non-stationary data streams with unknown distribution drift.
method Proposes split-conformal and full-conformal algorithms that adapt to drift detection and incorporate stability for online learning.
result Proves minimax-optimal regret for online full conformal algorithm under appropriate restrictions.
This guide simplifies high-probability regret bounds in empirical risk minimization.
problem High-probability regret bounds in empirical risk minimization.
method Modular presentation, three-step recipe, localized Rademacher complexity, local maximal inequalities, metric-entropy integrals.
result Recover familiar rates for various function classes and derive regret bounds for nuisance components.
This paper improves online learning algorithms for LP problems, achieving better regret bounds.
problem Achieving optimal regret bounds in online linear programming.
method Develops a new framework for first-order online learning algorithms under certain error bound conditions.
result First-order learning algorithms achieve o ( T ) o(\sqrt{T}) o ( T ) regret in continuous support and O ( log T ) \mathcal{O}(\log T) O ( log T ) regret in finite support, improving over O ( T ) \mathcal{O}(\sqrt{T}) O ( T ) . New algorithms minimize dynamic regret for strongly convex losses.
problem Minimizing dynamic regret for strongly convex losses.
method Developed Strongly Adaptive algorithms exploiting KKT conditions.
result Achieved near optimal dynamic regret of O ( d 1 / 3 n 1 / 3 e x t T V [ u 1 : n ] 2 / 3 ∨ d ) O(d^{1/3} n^{1/3} ext{TV}[u_{1:n}]^{2/3} \vee d) O ( d 1/3 n 1/3 e x t T V [ u 1 : n ] 2/3 ∨ d ) . New algorithm optimizes best arm identification with minimal regret.
problem Best arm identification in multi-armed bandit problems.
method Characterized Bayesian simple regret with continuity conditions of prior, proposed a simple algorithm.
result Proposed algorithm achieves rate-optimal Bayesian simple regret.
This paper analyzes regret bounds for Gaussian process Thompson sampling.
problem Analyzing the performance of Gaussian process Thompson sampling (GP-TS) in Bayesian optimization.
method The paper derives several regret bounds for GP-TS, including a lower bound, upper bounds on the second moment of cumulative regret, expected lenient regret, and improved cumulative regret.
result The paper provides improved regret upper bounds for GP-TS, showing that it suffers from a polynomial dependence on 1 / δ 1/δ 1/ δ with probability δ δ δ . We study combinatorial multi-armed bandit with probabilistically triggered arms (CMAB-T) and semi-bandit feedback. We resolve a serious issue in the prior CMAB-T studies where the regret bounds contain a possibly exponentially large factor of 1 / p ∗ 1/p^* 1/ p ∗ , where p ∗ p^* p ∗ is the minimum positive probability that an arm is trigg…
FP-UCB algorithm achieves bounded regret for finitely parameterized multi-armed bandits.
problem Finitely parameterized multi-armed bandits with unknown but known parameter set.
method FP-UCB algorithm using structural information about the parameter set.
result FP-UCB achieves bounded regret under structural condition, logarithmic otherwise.
New algorithm reduces regret in GLM bandits with tighter bounds.
problem Reducing regret in generalized linear contextual bandits.
method Double Doubly Robust (DDR) estimator for independence.
result First d \sqrt{d} d regret bound for GLM bandits. Develops a Best-of-Both-Worlds algorithm for linear contextual bandits with Tsallis entropy.
problem Linear contextual bandits with i.i.d. contexts.
method Follow-The-Regularized-Leader (FTRL) with Tsallis entropy.
result Achieves $O\left(\log(T)^{\frac{1+β}{2+β}}T^{\frac{1}{2+β}}
ight)$ regret under margin condition.
New algorithms minimize dynamic regret in non-stationary online learning.
problem Universal dynamic regret minimization under exp-concave and smooth losses.
method Strongly Adaptive algorithms with a path variational based on second order differences of the comparator sequence.
result Achieve a dynamic regret of i l d e O ( d 2 n 1 / 5 C n 2 / 5 ∨ d 2 ) ilde O(d^2 n^{1/5} C_n^{2/5} \vee d^2) i l d e O ( d 2 n 1/5 C n 2/5 ∨ d 2 ) , optimal modulo dependencies. Unified proof for various bandit algorithms with logarithmic regret.
problem Achieving logarithmic regret in stochastic bandit algorithms.
method Minimal high-probability concentration condition and two deterministic lemmas.
result Unified proofs for classical and contemporary bandit algorithms.
Paper addresses regret minimization and inference in high-dimensional online decision-making.
problem Regret minimization and statistical inference in high-dimensional online decision-making.
method Integrates ε-greedy bandit algorithm with hard thresholding for sparse bandit parameters and debiasing method for inference.
result Achieves either O ( T 1 / 2 ) O(T^{1/2}) O ( T 1/2 ) regret or O ( T 1 / 2 ) O(T^{1/2}) O ( T 1/2 ) -consistent inference, with trade-off between exploration and exploitation. Thresholded Lasso bandit minimizes regret in sparse linear bandits.
problem Sparse stochastic contextual linear bandits with large feature vectors.
method Uses Lasso framework with thresholding to estimate reward function and its sparse support.
result Non-asymptotic regret upper bounds scaling as O ( log d + T ) \mathcal{O}( \log d + \sqrt{T}) O ( log d + T ) . Study on deep neural networks for reward modeling with pairwise comparison data.
problem Reward modeling with deep neural networks in non-parametric settings.
method Established a non-asymptotic regret bound for deep reward estimators, introduced a margin-type condition.
result Improved regret bound for deep reward estimators, highlighting the importance of clear human beliefs.
New algorithm reduces best-in-class regret in contextual bandits.
problem Compete with the best policy in a class without model restrictions.
method Proposes an algorithm that updates policies by minimizing a pessimistic objective, including a clipped inverse-propensity estimate and variance penalty.
result Achieves fast best-in-class regret rates, including polylogarithmic rates in the parametric case.
Study optimal stopping for diffusion processes using data-driven methods.
problem Optimal stopping for diffusion processes under unknown conditions.
method Data-driven approach, deriving upper and lower bounds on simple and cumulative regret.
result Verified minimax optimality and improved convergence rates.
Develops a method to tackle high-dimensional linear bandits with knapsacks using online sparse estimation and dual variables.
problem High-dimensional linear bandits with knapsacks.
method Online hard thresholding algorithm for sparse estimation, integrated with primal-dual scheme.
result Achieves sub-linear regret that scales logarithmically with feature dimension, improving on prior work.
Study improves CTS's approximation regret for combinatorial bandits.
problem Improving CTS's performance on non-exact oracles.
method Develops a new O ( log ( T ) / Δ ) \mathcal{O}(\log(T)/Δ) O ( log ( T ) /Δ ) upper bound for CTS under specific conditions. result First O ( log ( T ) / Δ ) \mathcal{O}(\log(T)/Δ) O ( log ( T ) /Δ ) approximation regret upper bound for CTS. New betting strategy reduces regret to ln(ln n) with protection against adversarial data.
problem Tackles the problem of minimizing regret in betting against adversarial and stochastic data.
method Combines insights from Robbins and Cover, using a mixture strategy.
result Exhibits a regret of O(ln(ln n)) on almost all paths, with O(log n) regret on the complement.
Improved online learning for hidden-convex losses achieves optimal regret.
problem Adversarial online learning with nonconvex losses that become convex after reparameterization.
method Algorithmic equivalence between OGD and OMD on convex losses, with Hessian compatibility condition.
result OGD achieves O ( T ) \mathcal{O}(\sqrt{T}) O ( T ) regret for hidden-convex losses, matching optimal rate. New algorithm adapts to unknown demand smoothness for dynamic pricing.
problem Dynamic pricing with unknown Hölder smoothness of demand function.
method Self-similarity condition and adaptive algorithm.
result Adaptive algorithm achieves minimax optimal regret without prior knowledge of smoothness.
Paper stabilizes bandit learning with regularization, improving inference under adaptive sampling.
problem Challenges in statistical inference with adaptive sampling.
method Refined stability condition for online algorithms, using regularized stochastic-mirror-descent-style methods.
result Derives precise regret bounds and asymptotic normality, showing necessity of regularization for valid inference.
Adaptive designs achieve strong Neyman regret guarantees for ATE estimation.
problem Estimating unbiased average treatment effect in sequential experiments.
method Proposed adaptive designs with O ~ ( log T ) \widetilde{O}(\log T) O ( log T ) Neyman regret under boundedness assumptions and O ~ ( T ) \widetilde{O}(\sqrt{T}) O ( T ) multigroup Neyman regret in covariate-based settings. result Adaptive designs outperform non-adaptive designs in terms of Neyman regret, especially in covariate-based settings.
We present a generalization of the adversarial linear bandits framework, where the underlying losses are kernel functions (with an associated reproducing kernel Hilbert space) rather than linear functions. We study a version of the exponential weights algorithm and bound its regret in this setting. Under conditions on …
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.
TVBO optimizes time-varying functions with asymptotically vanishing regret.
problem Understanding the asymptotic performance of TVBO for time-varying black-box functions.
method Provided upper and lower bounds for cumulative regret of TVBO algorithms.
result TVBO algorithms can achieve asymptotically vanishing regret under certain conditions.
Online learning is a powerful tool for analyzing iterative algorithms. However, the classic adversarial setup sometimes fails to capture certain regularity in online problems in practice. Motivated by this, we establish a new setup, called Continuous Online Learning (COL), where the gradient of online loss function cha…
New algorithm achieves near-optimal performance in dueling bandit problem.
problem Optimizing decision-making in dueling bandit problems with limited adaptive rounds.
method Developed a batched algorithm that matches the asymptotic regret bounds of sequential algorithms under the Condorcet condition.
result Asymptotic regret of O ( K 2 log 2 ( K ) ) + O ( K log ( T ) ) O(K^2\log^2(K)) + O(K\log(T)) O ( K 2 log 2 ( K )) + O ( K log ( T )) in O ( log ( T ) ) O(\log(T)) O ( log ( T )) rounds. Paper analyzes FTPL's effectiveness in combinatorial semi-bandit problems.
problem Optimizing FTPL policy in combinatorial semi-bandit problems.
method Geometric resampling (GR) and conditional geometric resampling (CGR) for FTPL in semi-bandit setting.
result FTPL achieves optimal regret bounds in both Fréchet and Pareto distributions.
Sequential screening and dynamic regret in multi-armed bandits with arriving arms
problem Sequential experimentation with expanding arm set
method UCB-AA with preliminary screening
result Regret bounds depend on arrival process
New algorithm reduces dynamic regret for exp-concave losses.
problem Minimizing dynamic regret in online learning with exp-concave losses.
method Integrates KKT conditions to achieve optimal dynamic regret.
result Achieves dynamic regret of i l d e O ∗ ( n 1 / 3 C n 2 / 3 ∨ 1 ) ilde O^*(n^{1/3}C_n^{2/3} \vee 1) i l d e O ∗ ( n 1/3 C n 2/3 ∨ 1 ) . This paper improves Thompson Sampling for complex decision-making problems.
problem Learning in infinite-horizon discounted decision processes with unknown parameters.
method Developed a general canonical probability space and new metrics for analyzing adaptive learning algorithms.
result Thompson Sampling achieves complete learning in complex decision-making problems.
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.
The paper optimizes risk-sensitive RL with CVaR, achieving near-minimax-optimal results.
problem Optimizing risk-sensitive reinforcement learning with CVaR objective.
method Developed algorithms for multi-arm bandits and online RL in MDPs, achieving near-minimax-optimal regret.
result Achieved near-minimax-optimal regret of O ( τ − 1 S A K ) O(τ^{-1}\sqrt{SAK}) O ( τ − 1 S A K ) for constant τ τ τ . New method for semiparametric bandits reduces regret to optimal levels.
problem Complex reward structures in semiparametric bandits.
method Experimental-design approach with sharp regret bound and PAC bound.
result Minimax regret of i l d e O ( d T ) ilde{O}(\sqrt{dT}) i l d e O ( d T ) and logarithmic regret under positive suboptimality gap.