ICLR 2026 accepted papers on Reinforcement learning
390 of the 5,358 papers accepted at ICLR 2026 (7.3%) are on Reinforcement learning, compared with 5.8% at ICLR 2025. They fall into 6 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
RL from feedback · 134 papers
- Agentic Reinforcement Learning with Implicit Step Rewards (1 citation)
- COMAL: A Convergent Meta-Algorithm for Aligning LLMs with General Preferences (1 citation)
- CUDA-L1: Improving CUDA Optimization via Contrastive Reinforcement Learning (1 citation)
- KL-Regularized Reinforcement Learning is Designed to Mode Collapse (1 citation)
- Plan-R1: Safe and Feasible Trajectory Planning as Language Modeling (1 citation)
- ResT: Reshaping Token-Level Policy Gradients for Tool-Use Large Language Models (1 citation)
- The Art of Scaling Reinforcement Learning Compute for LLMs (1 citation)
- Adaptive Rollout Allocation for Online Reinforcement Learning with Verifiable Rewards
- All Roads Lead to Likelihood: The Value of Reinforcement Learning in Fine-Tuning
- AMPED: Adaptive Multi-objective Projection for balancing Exploration and skill Diversification
- ARM-FM: Automated Reward Machines via Foundation Models for Compositional Reinforcement Learning
- Asymmetric Proximal Policy Optimization: mini-critics boost LLM reasoning
and 122 more in the interactive map.
Visual and representation RL · 103 papers
- Learn the Ropes, Then Trust the Wins: Self-imitation with Progressive Exploration for Agentic Reinforcement Learning (1 citation)
- MobileRL: Online Agentic Reinforcement Learning for Mobile GUI Agents (1 citation)
- Ada-Diffuser: Latent-Aware Adaptive Diffusion for Decision-Making
- Automating the Refinement of Reinforcement Learning Specifications
- BFM-Zero: A Promptable Behavioral Foundation Model for Humanoid Control Using Unsupervised Reinforcement Learning
- Capturing Visual Environment Structure Correlates with Control Performance
- Chain-of-Context Learning: Dynamic Constraint Understanding for Multi-Task VRPs
- Compose Your Policies! Improving Diffusion-based or Flow-based Robot Policies via Test-time Distribution-level Composition
- ComputerRL: Scaling End-to-End Online Reinforcement Learning for Computer Use Agents
- Ctrl-World: A Controllable Generative World Model for Robot Manipulation
- D-REX: Differentiable Real-to-Sim-to-Real Engine for Learning Dexterous Grasping
- DataMIL: Selecting Data for Robot Imitation Learning with Datamodels
and 91 more in the interactive map.
Policy gradient and actor-critic · 62 papers
- Beyond Distributions: Geometric Action Control for Continuous Reinforcement Learning (1 citation)
- A Reward-Free Viewpoint on Multi-Objective Reinforcement Learning
- Accelerated Learning with Linear Temporal Logic using Differentiable Simulation
- Asynchronous Policy Gradient Aggregation for Efficient Distributed Reinforcement Learning
- CaRe-BN: Precise Moving Statistics for Stabilizing Spiking Neural Networks in Reinforcement Learning
- Chunking the Critic: A Transformer-based Soft Actor-Critic with N-Step Returns
- Decoupled Q-Chunking
- DiffusionNFT: Online Diffusion Reinforcement with Forward Process
- Distributional value gradients for stochastic environments
- Distributions as Actions: A Unified Framework for Diverse Action Spaces
- Does “Do Differentiable Simulators Give Better Policy Gradients?” Give Better Policy Gradients?
- Dual-Objective Reinforcement Learning with Novel Hamilton-Jacobi-Bellman Formulations
and 50 more in the interactive map.
Offline RL · 49 papers
- Bayes Adaptive Monte Carlo Tree Search for Offline Model-based Reinforcement Learning (1 citation)
- Accelerating Diffusion Planners in Offline RL via Reward-Aware Consistency Trajectory Distillation
- Action-Free Offline-To-Online RL via Discretised State Policies
- Algorithmic Guarantees for Distilling Supervised and Offline RL Datasets
- Belief-Based Offline Reinforcement Learning for Delay-Robust Policy Optimization
- Beyond Penalization: Diffusion-based Out-of-Distribution Detection and Selective Regularization in Offline Reinforcement Learning
- Cross-Embodiment Offline Reinforcement Learning for Heterogeneous Robot Datasets
- Deep SPI: Safe Policy Improvement via World Models
- DR-SAC: Distributionally Robust Soft Actor-Critic for Reinforcement Learning under Uncertainty
- Dual-Robust Cross-Domain Offline Reinforcement Learning Against Dynamics Shifts
- Efficient Reinforcement Learning by Guiding World Models with Non-Curated Data
- Enhancing Generative Auto-bidding with Offline Reward Evaluation and Policy Search
and 37 more in the interactive map.
Imitation learning · 21 papers
- Abstracting Robot Manipulation Skills via Mixture-of-Experts Diffusion Policies
- APC-RL: Exceeding data-driven behavior priors with adaptive policy composition
- CE-Nav: Flow-Guided Reinforcement Refinement for Cross-Embodiment Local Navigation
- GAR: Generative Adversarial Reinforcement Learning for Formal Theorem Proving
- Generative Adversarial Post-Training Mitigates Reward Hacking in Live Human-AI Music Interaction
- Geometry-aware Policy Imitation
- GRACE: A Language Model Framework for Explainable Inverse Reinforcement Learning
- Imitation Learning as Return Distribution Matching
- Learning to Answer from Correct Demonstrations
- Much Ado About Noising: Dispelling the Myths of Generative Robotic Control
- On Discovering Algorithms for Adversarial Imitation Learning
- Robustness in the Face of Partial Identifiability in Reward Learning
and 9 more in the interactive map.
MDPs and planning · 21 papers
- A Unifying View of Coverage in Linear Off-policy Evaluation
- An Orthogonal Learner for Individualized Outcomes in Markov Decision Processes
- AutoQD: Automatic Discovery of Diverse Behaviors with Quality-Diversity Optimization
- BRIDGE: Bi-level Reinforcement Learning for Dynamic Group Structure in Coalition Formation Games
- Flowing Through States: Neural ODE Regularization for Reinforcement Learning
- Near-Optimal Sample Complexity Bounds for Constrained Average-Reward MDPs
- Near-Optimal Second-Order Guarantees for Model-Based Adversarial Imitation Learning
- Partially Equivariant Reinforcement Learning in Symmetry-Breaking Environments
- Replicable Reinforcement Learning with Linear Function Approximation
- Solving General-Utility Markov Decision Processes in the Single-Trial Regime with Online Planning
- The Sample Complexity of Online Reinforcement Learning: A Multi-model Perspective
- Analysis of approximate linear programming solution to Markov decision problem with log barrier function
and 9 more in the interactive map.