NeurIPS 2026 accepted papers on Bandits and online learning
126 of the 6,080 papers accepted at NeurIPS 2026 (2.1%) are on Bandits and online learning, compared with 2.2% at NeurIPS 2025. They fall into 4 topics.
See NeurIPS 2026 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Online convex optimization and games · 67 papers
- A Memory Efficient Unified Algorithm for Online Learning of Linear Dynamical Systems
- A Reduction from Delayed to Immediate Feedback for Online Convex Optimization with Improved Guarantees
- A second order regret bound for NormalHedge
- Capacity-Constrained Online Convex Optimization with Delayed Feedback
- Distributed Online Convex Optimization with Compressed Communication: Optimal Regret and Applications
- Dynamic Regret in Online Convex Optimization with Indicator Switching Costs
- From Average Sensitivity to Small-Loss Regret Bounds under Random-Order Model
- From Non-Convex to Strongly Convex: Curvature-Adaptive FTPL for Online Optimization
- Instance-Adaptive Online Multicalibration
- Learning Augmented Exact Exponential Algorithms
- Learning to Persuade a Biased Receiver
- Near-Optimal Last-Iterate Convergence for Zero-Sum Games with Bandit Feedback and Opponent Actions
and 55 more in the interactive map.
Multi-armed bandits · 27 papers
- Achieving adaptivity and optimality for multi-armed bandits using Exponential-Kullback Leibler Maillard Sampling
- Batched Stochastic Linear Bandits with 1-Bit Communication Constraints
- Best Arm Identification in Generalized Linear Bandits via Hybrid Feedback
- Does 1/2-Tsallis-INF Also Work Well for Best-Arm Identification?
- Multi-Armed Bandits With Best-Action Queries
- Near-Optimal Stochastic Linear Bandits with Delay
- Nearly Optimal Fixed-Confidence Best-Arm Identification with 1-Bit Feedback
- On the Optimal Sample Complexity of Offline Multi-Armed Bandits with KL Regularization
- Practical Adversarial Attacks on Stochastic Bandits via Fake Data Injection
- Shuffle and Joint Differential Privacy for Generalized Linear Contextual Bandits
- The Sample Complexity of Multiple Change Point Identification under Bandit Feedback
- The Sampling Complexity of Condorcet Winner Identification in Dueling Bandits
and 15 more in the interactive map.
Contextual bandits and Thompson sampling · 23 papers
- Active Context Selection Improves Simple Regret in Contextual Bandits
- Algorithm for Contextual Queueing Bandits with Rate-Optimal Queue Length Regret
- Bandits via Additive Quantized Representations
- Block Optimism for Nonstationary Bandits with Latent Linear Dynamics
- Certified Policy Optimisation for Nested Causal Bandits via PAC-Bayes Risk
- Correlation-Aware Contextual Bandits with Surrogate Rewards for LLM Routing
- Direction-Aware Offline-to-Online Learning in Linear Contextual Bandits
- Improved Regret Analysis For Parallel Gaussian Process Bandit Optimization
- Learning to target with network interference
- Learning with Multiple Correct Answers - Regret Bounds under Different Feedback Models
- Lipschitz Dueling Bandits over Continuous Action Spaces
- Logistic Bandits with $\tilde{O}(\sqrt{dT})$ Regret without Context Diversity Assumptions
and 11 more in the interactive map.
RL theory and MDPs · 9 papers
- DARLING: Detection Augmented Reinforcement Learning with Non-Stationary Guarantees
- Information-Directed Offline-to-Online Reinforcement Learning
- Learning Weakly Communicating Average-Reward CMDPs: Strong Duality and Improved Regret
- Reinforcement Learning with Multi-Step Lookahead Information Via Adaptive Batching
- Lyapunov-Driven Optimistic Learning for Online Scheduling with Multi-Stage Tasks
- OPSRL-SSP: Optimistic Posterior Sampling for Stochastic Shortest Path with Minimax-Optimal Regret
- Personalized and Collaborative Online LQR via Thompson Sampling
- Policy Optimization in Tabular MDPs: Data-dependent Regret under Unknown Transitions
- Towards instance-dependent regret optimality in Episodic MDPs with Posterior Sampling