ICLR 2026 accepted papers on Bandits and online learning
52 of the 5,358 papers accepted at ICLR 2026 (1.0%) are on Bandits and online learning, compared with 1.0% at ICLR 2025. They fall into 4 topics.
See ICLR 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 · 23 papers
- A New Approach to Controlling Linear Dynamical Systems
- Better Bounds for the Distributed Experts Problem
- Convergence of Regret Matching in Potential Games and Constrained Optimization
- Discounted Online Convex Optimization: Uniform Regret Across a Continuous Interval
- Learning a Game by Paying the Agents
- Learning to Play Multi-Follower Bayesian Stackelberg Games
- Nearly-Optimal Bandit Learning in Stackelberg Games with Side Information
- Nonparametric Contextual Online Bilateral Trade
- Online Decision-Focused Learning
- Online Learning and Equilibrium Computation with Ranking Feedback
- Online Prediction of Stochastic Sequences with High Probability Regret Bounds
- Oracle-efficient Hybrid Learning with Constrained Adversaries
and 11 more in the interactive map.
Multi-armed bandits · 11 papers
- Bandits with Single-Peaked Preferences and Limited Resources
- Best-of-Majority: Minimax-Optimal Strategy for Pass@k Inference Scaling
- Combinatorial Rising Bandits
- Efficient Multi-objective Prompt Optimization via Pure-exploration Bandits
- In-Context Learning for Pure Exploration
- Lipschitz Bandits with Stochastic Delayed Feedback
- Multi-Armed Bandits with Minimum Aggregated Revenue Constraints
- Non-Asymptotic Analysis of (Sticky) Track-and-Stop
- A Near-Optimal Best-of-Both-Worlds Algorithm for Federated Bandits
- Best-of-three-worlds Analysis for Dueling Bandits with Borda Winner
- Efficient Adversarial Attacks on High-dimensional Offline Bandits
Contextual bandits and Thompson sampling · 10 papers
- Breaking the Total Variance Barrier: Sharp Sample Complexity for Linear Heteroscedastic Bandits with Fixed Action Set
- Diversified Multinomial Logit Contextual Bandits
- Laplacian Kernelized Bandit
- Online Minimization of Polarization and Disagreement via Low-Rank Matrix Bandits
- Queue Length Regret Bounds for Contextual Queueing Bandits
- Revisiting Matrix Sketching in Linear Bandits: Achieving Sublinear Regret via Dyadic Block Sketching
- Single Index Bandits: Generalized Linear Contextual Bandits with Unknown Reward Functions
- Contextual Causal Bayesian Optimisation
- Efficient Best-of-Both-Worlds Algorithms for Contextual Combinatorial Semi-Bandits
- Variance-Dependent Regret Lower Bounds for Contextual Bandits
RL theory and MDPs · 8 papers
- An Improved Model-free Decision-estimation Coefficient with Applications in Adversarial MDPs
- Frozen Policy Iteration: Computationally Efficient RL under Linear $Q^{\pi}$ Realizability for Deterministic Dynamics
- Is Pure Exploitation Sufficient in Exogenous MDPs with Linear Function Approximation?
- Learning in Prophet Inequalities with Noisy Observations
- Q-Learning with Fine-Grained Gap-Dependent Regret
- Q-learning with Posterior Sampling
- Minimax Optimal Adversarial Reinforcement Learning
- Primal-Dual Policy Optimization for Adversarial Linear CMDPs