Reinforcement Learning

Also known as RL

Ask about this topic

Overview

Reinforcement learning studies how an agent can choose actions over time to earn rewards, using experience to improve its decisions. Common approaches include value-based methods such as Q-learning, policy-gradient methods such as PPO, and model-based methods that learn or use a model of the environment; offline and hierarchical methods address learning from fixed data and long tasks. Applications range from Atari and MuJoCo benchmarks to robot control, traffic, games, and language-model agents that use tools or solve multi-step problems. A central challenge is making learning efficient and reliable when rewards are sparse or imperfect, environments change, and success in simulation or benchmarks may not transfer safely to deployment.

Latest papers

CardsList

1,742 papers237 papers in the last four weeks, up 46% on the four weeks before. 1.4% of all new papers.

Mar 6, 2026

Artificial Intelligence for Climate Adaptation: Reinforcement Learning for Climate Change-Resilient Transport

Climate change is expected to intensify rainfall and, consequently, pluvial flooding, leading to increased disruptions in urban transportation systems over the coming decades. Designing effective adaptation strategies is challenging due to the long-term, sequential nature of infrastructure investments, deep climate uncertainty, and the complex interactions between flooding, infrastructure, and mobility impacts. In this work, we propose a novel decision-support framework using reinforcement learning (RL) for long-term flood adaptation planning. Formulated as an integrated assessment model (IAM), the framework combines rainfall projection and flood modeling, transport simulation, and quantification of direct and indirect impacts on infrastructure and mobility. Our RL-based approach learns adaptive strategies that balance investment and maintenance costs against avoided impacts. We evaluate the framework through a case study of Copenhagen's inner city over the 2024-2100 period, testing multiple adaptation options, and different belief and realized climate scenarios. Results show that the framework outperforms traditional optimization approaches by discovering coordinated spatial and temporal adaptation pathways and learning trade-offs between impact reduction and adaptation investment, yielding more resilient strategies. Overall, our results showcase the potential of reinforcement learning as a flexible decision-support tool for adaptive infrastructure planning under climate uncertainty.

Miguel Costa, Arthur Vandervoort, Carolin Schmidt +5

Department of Technology, Management and Economics, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark · Present address: School of Management, Technical University of Munich, 80333 Munich, Germany. · Present address: Saint-Gobain Distribution Denmark, 2605 Brøndby, Denmark. +1

Reinforcement Learning
AskPDF
Mar 6, 2026

Synthetic Monitoring Environments for Reinforcement Learning

Reinforcement Learning (RL) lacks benchmarks that enable precise, white-box diagnostics of agent behavior. Current environments often entangle complexity factors and lack ground-truth optimality metrics, making it difficult to isolate why algorithms fail. We introduce Synthetic Monitoring Environments (SMEs), an infinite suite of continuous control tasks. SMEs provide fully configurable task characteristics and known optimal policies. As such, SMEs allow for the exact calculation of instantaneous regret. Their rigorous geometric state space bounds allow for systematic within-distribution (WD) and out-of-distribution (OOD) evaluation. We demonstrate the framework's benefit through multidimensional ablations of PPO, TD3, and SAC, revealing how specific environmental properties - such as action or state space size, reward sparsity and complexity of the optimal policy - impact WD and OOD performance. We thereby show that SMEs offer a standardized, transparent testbed for transitioning RL evaluation from empirical benchmarking toward rigorous scientific analysis.

Leonard Pleiss, Carolin Schmidt, Maximilian Schiffer

Technical University Munich Munich, 80331

RL BenchmarksReinforcement LearningGeneralization in RLContinuous Control
AskPDF
Mar 6, 2026

Stabilizing Reinforcement Learning for Diffusion Language Models

Group Relative Policy Optimization (GRPO) is highly effective for post-training autoregressive (AR) language models, yet its direct application to diffusion large language models (dLLMs) often triggers reward collapse. We identify two sources of incompatibility. First, GRPO relies on importance ratios defined by sequence probabilities, which are intractable in dLLMs and must be estimated (e.g., via ELBO-based or mean-field likelihood proxies), yielding inherently noisy ratios. Second, standard GRPO's formulation is not designed for estimated ratios: its conditional clipping can be anomalously bypassed by model-agnostic estimation noise, producing gradient spikes, while its fixed group-size normalization amplifies gradient-magnitude fluctuations under high-variance ratio estimates. We show these effects form a self-reinforcing instability loop that drives policy drift and further increases ratio variance. To break this loop, we propose StableDRL, a reformulation of GRPO tailored for dLLMs that uses (i) unconditional clipping to suppress outlier-induced spikes and (ii) self-normalization to constrain updates within the convex hull of per-sample gradients. We further extend StableDRL to block-wise diffusion models via a staircase attention mechanism.

Jianyuan Zhong, Kaibo Wang, Ding Ding +5

Huawei Foundation Model Department · The Chinese University of Hong Kong · The Hong Kong University of Science and Technology

RL for Diffusion ModelsReinforcement LearningGroup Relative Policy Optimization
AskPDF
Mar 6, 2026

Preventing Learning Stagnation in PPO by Scaling to 1 Million Parallel Environments

An agent's performance stagnating at a suboptimal level is a common problem in deep on-policy RL. Focusing on PPO, we show that plateaus in certain regimes arise not because of known exploration, capacity, or optimisation challenges, but because sample-based estimates of the loss eventually become poor proxies for the true objective over the course of training. Looking deeper, PPO alternates between sampling rollouts from several parallel environments online using the current policy (which we call the "outer loop") and performing repeated minibatch SGD steps against this offline dataset (the "inner loop"). In our work, we abstract away the inner loop, and conceptually model the outer loop as standard stochastic optimisation. The step size is then controlled by the regularisation strength towards the previous policy and the gradient noise by the number of samples collected between policy update steps. This framing predicts that, much like in SGD, if the outer step size is too large relative to the noise, updates become uninformative and lead to the policy thrashing around a local optimum instead of converging. Recasting PPO in this light makes it clear that there are two ways to address this particular type of learning stagnation: either reduce the step size or increase the number of samples collected between updates. We validate the predictions of our model and conclude that increasing the number of parallel environments is a simple way to avoid these plateaus by simultaneously altering both these factors. Applying our analysis and scaling PPO to more than 1M parallel environments enables monotonic performance improvement up to one trillion transitions and leads to vastly superior performance compared to prior baselines in a complex open-ended domain.

Michael Beukman, Khimya Khetarpal, Zeyu Zheng +4

FLAIR, University of Oxford · Google DeepMind

Reinforcement LearningProximal Policy Optimization
AskPDF
Mar 5, 2026

Reinforcement Learning for Power-Flow Network Analysis

The power flow equations are non-linear multivariate equations that describe the relationship between power injections and bus voltages of electric power networks. Given a network topology, we are interested in finding network parameters with many equilibrium points. This corresponds to finding instances of the power flow equations with many real solutions. Current state-of-the art algorithms in computational algebra are not capable of answering this question for networks involving more than a small number of variables. To remedy this, we design a probabilistic reward function that gives a good approximation to this root count, and a state-space that mimics the space of power flow equations. We derive the average root count for a Gaussian model, and use this as a baseline for our RL agents. The agents discover instances of the power flow equations with many more solutions than the average baseline. This demonstrates the potential of RL for power-flow network design and analysis as well as the potential for RL to contribute meaningfully to problems that involve complex non-linear algebra or geometry. \footnote{Author order alphabetic, all authors contributed equally.

Alperen Ergur, Julia Lindberg, Vinny Miller

The University of Texas at San Antonio, Mathematics and Computer Science Departments · Georgia Tech, Mathematics Department · The University of Texas at San Antonio, Computer Science Department

Power SystemsReinforcement Learning
AskPDF
Mar 5, 2026

A Novel Hybrid Heuristic-Reinforcement Learning Optimization Approach for a Class of Railcar Shunting Problems

Railcar shunting is a core planning task in freight railyards, where yard planners need to disassemble and reassemble groups of railcars to form outbound trains. Classification tracks with access from one side only can be considered as stack structures, where railcars are added and removed from only one end, leading to a last-in-first-out (LIFO) retrieval order. In contrast, two-sided tracks function like queue structures, allowing railcars to be added from one end and removed from the opposite end, following a first-in-first-out (FIFO) order. We consider a problem requiring assembly of multiple outbound trains using two locomotives in a railyard with two-sided classification track access. To address this combinatorially challenging problem class, we decompose the problem into two subproblems, each with one-sided classification track access and a locomotive on each side. We present a novel Hybrid Heuristic-Reinforcement Learning (HHRL) framework that integrates railway-specific heuristic solution approaches with a reinforcement learning method, specifically Q-learning. The proposed framework leverages methods to decrease the state-action space and guide exploration during reinforcement learning. The results of a series of numerical experiments demonstrate the efficiency and quality of the HHRL algorithm in both one-sided access, single-locomotive problems and two-sided access, two-locomotive problems.

Ruonan Zhao, Joseph Geunes

Wm Michael Barnes ’64 Department of Industrial and Systems Engineering, Texas A&M University

RL for Combinatorial OptimizationReinforcement Learning
AskPDF
Mar 5, 2026

Wiki-R1: Incentivizing Multimodal Reasoning for Knowledge-based VQA via Data and Sampling Curriculum

Knowledge-Based Visual Question Answering (KB-VQA) requires models to answer questions about an image by integrating external knowledge, posing significant challenges due to noisy retrieval and the structured, encyclopedic nature of the knowledge base. These characteristics create a distributional gap from pretrained multimodal large language models (MLLMs), making effective reasoning and domain adaptation difficult in the post-training stage. In this work, we propose Wiki-R1, a data-generation-based curriculum reinforcement learning framework that systematically incentivizes reasoning in MLLMs for KB-VQA. Wiki-R1 constructs a sequence of training distributions aligned with the model's evolving capability, bridging the gap from pretraining to the KB-VQA target distribution. We introduce controllable curriculum data generation, which manipulates the retriever to produce samples at desired difficulty levels, and a curriculum sampling strategy that selects informative samples likely to yield non-zero advantages during RL updates. Sample difficulty is estimated using observed rewards and propagated to unobserved samples to guide learning. Experiments on two KB-VQA benchmarks, Encyclopedic VQA and InfoSeek, demonstrate that Wiki-R1 achieves new state-of-the-art results, improving accuracy from 35.5% to 37.1% on Encyclopedic VQA and from 40.1% to 44.1% on InfoSeek. The project page is available at https://artanic30.github.io/project_pages/WikiR1/.

Shan Ning, Longtian Qiu, Xuming He

ShanghaiTech University, Shanghai, China · Lingang Laboratory, Shanghai, China · Shanghai Engineering Research Center of Intelligent Vision and Imaging

Visual Question AnsweringReinforcement LearningCurriculum RLVLM ReasoningKnowledge-Based VQARL for VLMs
AskPDF
Mar 5, 2026

Decoupling Task and Behavior: A Two-Stage Reward Curriculum in Reinforcement Learning for Robotics

Deep Reinforcement Learning is a promising tool for robotic control, yet practical application is often hindered by the difficulty of designing effective reward functions. Real-world tasks typically require optimizing multiple objectives simultaneously, necessitating precise tuning of their weights to learn a policy with the desired characteristics. To address this, we propose a two-stage reward curriculum where we decouple task-specific objectives from behavioral terms. In our method, we first train the agent on a simplified task-only reward function to ensure effective exploration before introducing the full reward that includes auxiliary behavior-related terms such as energy efficiency. Further, we analyze various transition strategies and demonstrate that reusing samples between phases is critical for training stability. We validate our approach on the DeepMind Control Suite, ManiSkill3, and a mobile robot environment, modified to include auxiliary behavioral objectives. Our method proves to be simple yet effective, substantially outperforming baselines trained directly on the full reward while exhibiting higher robustness to specific reward weightings.

Kilian Freitag, Knut Åkesson, Morteza Haghir Chehreghani

Chalmers University of Technology · Gothenburg University

RL for RoboticsReinforcement LearningReward Design for RLCurriculum RLMulti-Objective Reinforcement Learning
AskPDF
Mar 5, 2026

Reward-Conditioned Reinforcement Learning

Single-task RL agents are typically trained under a fixed reward function, which limits their robustness to reward misspecification and their ability to adapt to changing preferences. We introduce Reward-Conditioned Reinforcement Learning (RCRL), an off-policy method that conditions agents on reward parameterizations while collecting experience under a single nominal objective. By recomputing counterfactual rewards from shared replay data, RCRL exposes the agent to multiple reward objectives without additional environment interaction, connecting single-task RL with ideas from multi-objective and multi-task learning. Across single-task, multi-task, and vision-based benchmarks, RCRL improves sample efficiency under the nominal reward parameterization, enables efficient adaptation to new parameterizations, and supports zero-shot behavioral adjustment at deployment. Our results show that RCRL provides a scalable mechanism for learning robust, steerable policies without sacrificing the simplicity of single-task training.

Michal Nauman, Marek Cygan, Pieter Abbeel

University of Warsaw · University of Warsaw, Nomagic · UC Berkeley, Amazon FAR

Reinforcement LearningOff-Policy RLMulti-Objective Reinforcement Learning
AskPDF
Mar 4, 2026

Prism: Policy Reuse via Interpretable Strategy Mapping in Reinforcement Learning

We present PRISM (Policy Reuse via Interpretable Strategy Mapping), a framework that grounds reinforcement learning agents' decisions in discrete, causally validated concepts and uses those concepts as a zero-shot transfer interface between agents trained with different algorithms. PRISM clusters each agent's encoder features into KK concepts via K-means. Causal intervention establishes that these concepts directly drive - not merely correlate with - agent behavior: overriding concept assignments changes the selected action in 69.4% of interventions (p=8.6×10−86p = 8.6 \times 10^{-86}, 2500 interventions). Concept importance and usage frequency are dissociated: the most-used concept (C47, 33.0% frequency) causes only a 9.4% win-rate drop when ablated, while ablating C16 (15.4% frequency) collapses win rate from 100% to 51.8%. Because concepts causally encode strategy, aligning them via optimal bipartite matching transfers strategic knowledge zero-shot. On Go~7×\times7 with three independently trained agents, concept transfer achieves 69.5%±\pm3.2% and 76.4%±\pm3.4% win rate against a standard engine across the two successful transfer pairs (10 seeds), compared to 3.5% for a random agent and 9.2% without alignment. Transfer succeeds when the source policy is strong; geometric alignment quality predicts nothing (R2≈0R^2 \approx 0). The framework is scoped to domains where strategic state is naturally discrete: the identical pipeline on Atari Breakout yields bottleneck policies at random-agent performance, confirming that the Go results reflect a structural property of the domain.

Thomas Pravetz

RL for Game AIZero-Shot Policy TransferReinforcement LearningTransfer LearningInterpretability in RL
AskPDF
Mar 4, 2026

What Does Flow Matching Bring To TD Learning?

Recent work shows that flow matching can be effective for scalar Q-value function estimation in reinforcement learning (RL), but it remains unclear why or how this approach differs from standard critics. Contrary to conventional belief, we show that their success is not explained by distributional RL, as explicitly modeling return distributions can reduce performance. Instead, we argue that the use of integration for reading out values and dense velocity supervision at each step of this integration process for training improves TD learning via two mechanisms. First, it enables robust value prediction through test-time recovery, whereby iterative computation through integration dampens errors in early value estimates as more integration steps are performed. This recovery mechanism is absent in monolithic critics. Second, supervising the velocity field at multiple interpolant values induces more plastic feature learning within the network, allowing critics to represent non-stationary TD targets without discarding previously learned features or overfitting to individual TD targets encountered during training. We formalize these effects and validate them empirically, showing that flow-matching critics substantially outperform monolithic critics (2×\times in final performance and around 5×\times in sample efficiency) in settings where loss of plasticity poses a challenge e.g., in high-UTD online RL problems, while remaining stable during learning.

Bhavya Agrawalla, Michal Nauman, Aviral Kumar

Carnegie Mellon University · University of Warsaw

Reinforcement LearningValue Function EstimationTemporal-Difference Learning
AskPDF
Mar 4, 2026

Unbiased Dynamic Pruning for Efficient Group-Based Policy Optimization

Group Relative Policy Optimization (GRPO) effectively scales LLM reasoning but incurs prohibitive computational costs due to its extensive group-based sampling requirement. While recent selective data utilization methods can mitigate this overhead, they could induce estimation bias by altering the underlying sampling distribution, compromising theoretical rigor and convergence behavior. To address this limitation, we propose Dynamic Pruning Policy Optimization (DPPO), a framework that enables dynamic pruning while preserving unbiased gradient estimation through importance sampling-based correction. By incorporating mathematically derived rescaling factors, DPPO significantly accelerates GRPO training without altering the optimization objective of the full-batch baseline. Furthermore, to mitigate the data sparsity induced by pruning, we introduce Dense Prompt Packing, a window-based greedy strategy that maximizes valid token density and hardware utilization. Extensive experiments demonstrate that DPPO consistently accelerates training across diverse models and benchmarks. For instance, on Qwen3-4B trained on MATH, DPPO achieves 2.37×\times training speedup and outperforms GRPO by 3.36% in average accuracy across six mathematical reasoning benchmarks.

Haodong Zhu, Yangyang Ren, Yanjing Li +6

Beihang University · Zhongguancun Academy · Rakuten Singapore +3

Reinforcement LearningRL for Language Model ReasoningGroup Relative Policy Optimization
AskPDF
Mar 4, 2026

BeamPERL: Parameter-Efficient RL with Verifiable Rewards Specializes Compact LLMs for Structured Beam Mechanics Reasoning

Can reinforcement learning with hard, verifiable rewards teach a compact language model to reason about physics, or does it primarily learn to pattern-match toward correct answers? We study this question by training a 1.5B-parameter reasoning model on beam statics, a classic engineering problem, using parameter-efficient RLVR with binary correctness rewards from symbolic solvers, without teacher-generated reasoning traces. The best BeamPERL checkpoint achieves a 66.7% improvement in Pass@1 over the base model. However, the learned competence is anisotropic: the model generalizes compositionally (more loads) but fails under topological shifts (moved supports) that require the same equilibrium equations. Intermediate checkpoints yield the strongest reasoning, while continued optimization degrades robustness while maintaining reward. These findings reveal a key limitation of outcome-level alignment: reinforcement learning with exact physics rewards induces procedural solution templates rather than internalization of governing equations. The precision of the reward signal - even when analytically exact - does not by itself guarantee transferable physical reasoning. Our results suggest that verifiable rewards may need to be paired with structured reasoning scaffolding to move beyond template matching toward robust scientific reasoning.

Tarjei Paule Hage, Markus J. Buehler

Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA · Department of Civil and Environmental Engineering, Department of Mechanical Engineering, Schwarzman College of Computing, Massachusetts Institute of Technology, Cambridge, MA, USA

Structural EngineeringRL for Language Model ReasoningScientific Reasoning in Language ModelsReinforcement Learning with Verifiable RewardsReinforcement LearningLanguage Model Robustness
AskPDF
Mar 4, 2026

GIPO: Gaussian Importance Sampling Policy Optimization

Post-training with reinforcement learning (RL) has recently shown strong promise for advancing multimodal agents beyond supervised imitation. However, RL remains limited by poor data efficiency, particularly in settings where interaction data are scarce and quickly become outdated. To address this challenge, GIPO (Gaussian Importance sampling Policy Optimization) is proposed as a policy optimization objective based on truncated importance sampling, replacing hard clipping with a log-ratio-based Gaussian trust weight to softly damp extreme importance ratios while maintaining non-zero gradients. Theoretical analysis shows that GIPO introduces an implicit, tunable constraint on the update magnitude, while concentration bounds guarantee robustness and stability under finite-sample estimation. Experimental results show that GIPO achieves state-of-the-art performance among clipping-based baselines across a wide range of replay buffer sizes, from near on-policy to highly stale data, while exhibiting superior bias--variance trade-off, high training stability and improved sample efficiency. Code is available at https://github.com/distanceLu/GIPO.

Chengxuan Lu, Zhenquan Zhang, Shukuan Wang +3

1IROOTECH TECHNOLOGY · 2Wolf 1069 b Lab, Sany Group

Reinforcement LearningImportance Sampling for RLRL Post-TrainingSample-Efficient RLOff-Policy RL
AskPDF
Mar 4, 2026

Learning Approximate Nash Equilibria in Cooperative Multi-Agent Reinforcement Learning via Mean-Field Subsampling

Many large-scale platforms and networked control systems have a centralized decision maker interacting with a massive population of agents under strict observability constraints. Motivated by such applications, we study a cooperative Markov game with a global agent and nn homogeneous local agents in a communication-constrained regime, where the global agent only observes a subset of kk local agent states per time step. We propose an alternating learning framework (ALTERNATING-MARL)(\texttt{ALTERNATING-MARL}), where the global agent performs subsampled mean-field QQ-learning against a fixed local policy, and local agents update by optimizing in an induced MDP. We prove that these approximate best-response dynamics converge to an O~(1/k)\widetilde{O}(1/\sqrt{k})-approximate Nash Equilibrium, while separating the sample complexities between the joint state and action spaces. Finally, we validate our results in numerical simulations for multi-robot control.

Emile Anand, Ishani Karmarkar

Georgia Institute of Technology, Atlanta, GA · Stanford University, Palo Alto, CA

Mean-Field MARLReinforcement LearningCooperative MARLMulti-Agent CoordinationPartially Observable RL
AskPDF
Mar 4, 2026

Alternating Reinforcement Learning with Contextual Rubric Rewards: Beyond the Scalarization Strategy

Reinforcement Learning with Rubric Rewards (RLRR) is a framework that extends conventional reinforcement learning from human feedback (RLHF) and verifiable rewards (RLVR) by replacing scalar preference signals with structured, multi-dimensional, contextual rubric-based evaluations. However, existing approaches in RLRR are limited to linearly compressing vector rewards into a scalar reward with a fixed weightings, which is sensitive to artificial score design and fails to capture correlations among reward dimensions. To overcome the limitations of reward aggregation, this work proposes Alternating Reinforcement Learning with Rubric Rewards (ARL-RR), a framework that eliminates the need for a fixed scalarization by optimizing one semantic rubric meta-class at a time. Theoretically, we show that reward aggregation induces a variance contraction effect, which helps explain the performance gains. We further introduce a lightweight, search-based adaptation procedure that selects the next meta-class dynamically based on task performance, enabling the policy to emphasize critical objectives and thereby improve the model performance. Empirically, our experiments on the HealthBench dataset with experts annotations demonstrate that ARL-RR uniformly outperforms scalarized methods in both model performance and training efficiency across different model scales (1.7B, 4B, 8B, and 14B).

Guangchen Lan, Lian Xiong, Xin Zhou +7

Amazon · University of California, San Diego

Reinforcement LearningRubric-Based RLMulti-Objective Reinforcement Learning
AskPDF
Mar 4, 2026

Freezing of Gait Prediction using Proactive Agent that Learns from Selected Experience and DDQN Algorithm

Freezing of Gait (FOG) is a debilitating motor symptom commonly experienced by individuals with Parkinson's Disease (PD) which often leads to falls and reduced mobility. Timely and accurate prediction of FOG episodes is essential for enabling proactive interventions through assistive technologies. This study presents a reinforcement learning-based framework designed to identify optimal pre-FOG onset points, thereby extending the prediction horizon for anticipatory cueing systems. The model implements a Double Deep Q-Network (DDQN) architecture enhanced with Prioritized Experience Replay (PER) allowing the agent to focus learning on high-impact experiences and refine its policy. Trained over 9000 episodes with a reward shaping strategy that promotes cautious decision-making, the agent demonstrated robust performance in both subject-dependent and subject-independent evaluations. The model achieved a prediction horizon of up to 8.72 seconds prior to FOG onset in subject-independent scenarios and 7.89 seconds in subject-dependent settings. These results highlight the model's potential for integration into wearable assistive devices, offering timely and personalized interventions to mitigate FOG in PD patients.

Septian Enggar Sukmana, Sang Won Bae, Tomohiro Shibata

Kyushu Institute of Technology · Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology, Kitakyushu, Japan · Stevens Institute of Technology +1

Prioritized Experience ReplayParkinson's DiseaseGait AnalysisReinforcement Learning
AskPDF
Mar 4, 2026

Hybrid Belief Reinforcement Learning for Efficient Coordinated Spatial Exploration

Coordinating multiple autonomous agents to explore and serve spatially heterogeneous demand requires jointly learning unknown spatial patterns and planning trajectories that maximize task performance. Pure model-based approaches provide structured uncertainty estimates but lack adaptive policy learning, while deep reinforcement learning often suffers from poor sample efficiency when spatial priors are absent. This paper presents a hybrid belief-reinforcement learning (HBRL) framework to address this gap. In the first phase, agents construct spatial beliefs using a Log-Gaussian Cox Process (LGCP) and execute information-driven trajectories guided by a Pathwise Mutual Information (PathMI) planner with multi-step lookahead. In the second phase, trajectory control is transferred to a Soft Actor-Critic (SAC) agent, warm-started through dual-channel knowledge transfer: i) belief state transfer supplies spatial uncertainty, and ii) replay buffer seeding provides demonstration trajectories generated during LGCP exploration. A variance-normalized overlap penalty enables coordinated coverage through shared belief state, permitting cooperative sensing in high-uncertainty regions while discouraging redundant coverage in well-explored areas. The framework is instantiated in two scenarios: multi-Unmanned Aerial Vehicle (UAV) wireless service provisioning, and mobility-on-demand sensing over measured New York City taxi demand. In the wireless scenario it achieves 10.8% higher cumulative reward and takes 38% fewer episodes to reach the reward level of the baselines, with ablation studies confirming that dual-channel transfer outperforms either channel alone; the taxi scenario yields consistent gains with all algorithmic components unchanged.

Danish Rizvi, David Boyle

Imperial College London, UK

UAV CommunicationReinforcement LearningMulti-Robot ExplorationCooperative MARLBelief-Space Planning
AskPDF
Mar 3, 2026

Q-Measure-Learning for Continuous State RL: Efficient Implementation and Convergence

We study reinforcement learning in infinite-horizon discounted Markov decision processes with continuous state spaces, where data are generated online from a single trajectory under a Markovian behavior policy. To avoid maintaining an infinite-dimensional, function-valued estimate, we propose the novel Q-Measure-Learning, which learns a signed empirical measure supported on visited state-action pairs and reconstructs an action-value estimate via kernel integration. The method jointly estimates the stationary distribution of the behavior chain and the Q-measure through coupled stochastic approximation, leading to an efficient weight-based implementation with O(n)O(n) memory and O(n)O(n) computation cost per iteration. Under uniform ergodicity of the behavior chain, we prove almost sure sup-norm convergence of the induced Q-function to the fixed point of a kernel-smoothed Bellman operator. We also bound the approximation error between this limit and the optimal Q∗Q^* as a function of the kernel bandwidth. To assess the performance of our proposed algorithm, we conduct RL experiments in a two-item inventory control setting.

Shengbo Wang

Daniel J. Epstein Department of Industrial and Systems Engineering, University of Southern California

Reinforcement LearningQ-LearningFunction Approximation in RL
AskPDF
Mar 3, 2026

Minimax Optimal Strategy for Delayed Observations in Online Reinforcement Learning

We study reinforcement learning with delayed state observation, where the agent observes the current state after some random number of time steps. We propose an algorithm that combines the augmentation method and the upper confidence bound approach. For tabular Markov decision processes (MDPs), we derive a regret bound of O~(HDmax⁡SAK)\tilde{\mathcal{O}}(H \sqrt{D_{\max} SAK}), where SS and AA are the cardinalities of the state and action spaces, HH is the time horizon, KK is the number of episodes, and Dmax⁡D_{\max} is the maximum length of the delay. We also provide a matching lower bound up to logarithmic factors, showing the optimality of our approach. Our analytical framework formulates this problem as a special case of a broader class of MDPs, where their transition dynamics decompose into a known component and an unknown but structured component. We establish general results for this abstract setting, which may be of independent interest.

Harin Lee, Kevin Jamieson

Paul G. Allen School of Computer Science & Engineering, University of Washington, Seattle, Washington.

Delayed RLReinforcement LearningRegret Minimization in RL
AskPDF
Mar 3, 2026

Edit in 2D, Verify in 3D: Reinforcement Learning for Multi-view Consistent Scene Editing

Leveraging the priors of 2D diffusion models for 3D editing has emerged as a promising paradigm. However, multi-view consistency remains challenging in edited results, and the extreme scarcity of paired 3D-consistent editing data makes supervised fine-tuning (SFT) impractical, despite its effectiveness for editing tasks. In this paper, we observe that, while generating multi-view consistent 3D content is highly challenging, verifying 3D consistency is tractable, naturally positioning reinforcement learning (RL) as a feasible solution. Motivated by this, we propose RL3DEdit, a single-pass framework driven by RL optimization with novel rewards derived from the 3D foundation model, VGGT. Specifically, we leverage VGGT's robust priors learned from massive real-world data, feed the edited images into it, and utilize the output confidence maps and pose estimation errors as reward signals, effectively anchoring the 2D editing priors onto a 3D-consistent manifold via RL. Extensive experiments demonstrate that RL3DEdit achieves stable multi-view consistency and outperforms state-of-the-art methods in editing quality with high efficiency. To promote the development of 3D editing, we will release the code and model.

Jiyuan Wang, Chunyu Lin, Lei Sun +8

1BJTU, 2AMap Alibaba Group, 3NTU, 4CQUPT

Multi-View ConsistencyReinforcement Learning3D Scene EditingDiffusion-Based Image EditingRL for Image Editing
AskPDF
Mar 3, 2026

Reinforcement Learning with Symbolic Reward Machines

Reward Machines (RMs) are an established mechanism in Reinforcement Learning (RL) to represent and learn sparse, temporally extended tasks with non-Markovian rewards. RMs rely on high-level information in the form of labels that are emitted by the environment alongside the observation. However, this concept requires manual user input for each environment and task. The user has to create a suitable labeling function that computes the labels. These limitations lead to poor applicability in widely adopted RL frameworks. We propose Symbolic Reward Machines (SRMs) together with the learning algorithms QSRM and LSRM to overcome the limitations of RMs. SRMs consume only the standard output of the environment and process the observation directly through guards that are represented by symbolic formulas. In our evaluation, our SRM methods outperform the baseline RL approaches and generate the same results as the existing RM methods. At the same time, our methods adhere to the widely used environment definition and provide interpretable representations of the task to the user.

Thomas Krug, Daniel Neider

TU Dortmund University · Research Center Trustworthy Data Science and Security

Non-Markovian RLReinforcement LearningSparse-Reward RL
AskPDF
Mar 3, 2026

Enhancing User Throughput in Multi-panel mmWave Radio Access Networks for Beam-based MU-MIMO Using a DRL Method

Millimeter-wave (mmWave) communication systems, particularly those leveraging multi-user multiple-input and multiple-output (MU-MIMO) with hybrid beamforming, face challenges in optimizing user throughput and minimizing latency due to the high complexity of dynamic beam selection and management. This paper introduces a deep reinforcement learning (DRL) approach for enhancing user throughput in multi-panel mmWave radio access networks in a practical network setup. Our DRL-based formulation utilizes an adaptive beam management strategy that models the interaction between the communication agent and its environment as a Markov decision process (MDP), optimizing beam selection based on real-time observations. The proposed framework exploits spatial domain (SD) characteristics by incorporating the cross-correlation between the beams in different antenna panels, the measured reference signal received power (RSRP), and the beam usage statistics to dynamically adjust beamforming decisions. As a result, the spectral efficiency is improved and end-to-end latency is reduced. The numerical results demonstrate an increase in throughput of up to 16% and a reduction in latency by factors 3-7x compared to baseline (legacy beam management).

Ramin Hashemi, Vismika Ranasinghe, Teemu Veijalainen +2

Nokia, Espoo, Finland · Department of Information and Communications Engineering, Aalto University, Espoo, Finland

Millimeter-Wave CommunicationsWireless CommunicationsWireless Resource AllocationReinforcement Learning
AskPDF
Mar 3, 2026

Real-Time Generative Policy via Langevin-Guided Flow Matching for Autonomous Driving

Reinforcement learning (RL) is a fundamental methodology in autonomous driving systems, where generative policies exhibit considerable potential by leveraging their ability to model complex distributions to enhance exploration. However, their inherent high inference latency severely impedes their deployment in real-time decision-making and control. To address this issue, we propose diffusion actor-critic with entropy regulator via flow matching (DACER-F) by introducing flow matching into online RL, enabling the generation of competitive actions in a single inference step. By leveraging Langevin dynamics and gradients of the Q-function, DACER-F dynamically optimizes actions from experience replay toward a target distribution that balances high Q-value information with exploratory behavior. The flow policy is then trained to efficiently learn a mapping from a simple prior distribution to this dynamic target. In complex multi-lane and intersection simulations, DACER-F outperforms baselines diffusion actor-critic with entropy regulator (DACER) and distributional soft actor-critic (DSAC), while maintaining an ultra-low inference latency. DACER-F further demonstrates its scalability on standard RL benchmark DeepMind Control Suite (DMC), achieving a score of 775.8 in the humanoid-stand task and surpassing prior methods. Collectively, these results establish DACER-F as a high-performance and computationally efficient RL algorithm.

Tianze Zhu, Yinuo Wang, Wenjun Zou +6

School of Vehicle and Mobility, Tsinghua University, Beijing, China · College of Artificial Intelligence, Tsinghua University, Beijing, China

Reinforcement LearningRL for Autonomous DrivingRL for Generative Models
AskPDF
Mar 2, 2026

COOL-MC: Verifying and Explaining RL Policies for Platelet Inventory Management

Platelets expire within five days. Blood banks face uncertain daily demand and must balance ordering decisions between costly wastage from overstocking and life-threatening shortages from understocking. Reinforcement learning (RL) can learn effective ordering policies for this Markov decision process (MDP), but the resulting neural policies remain black boxes, hindering trust and adoption in safety-critical domains. We apply COOL-MC, a tool that combines RL with probabilistic model checking and explainable RL, to verify and explain a trained policy for the MDP on platelet inventory management inspired by Haijema et al. By constructing a policy-induced discrete-time Markov chain (which includes only the reachable states under the trained policy to reduce memory usage), we verify PCTL properties and provide feature-level explanations. Results show that the trained policy achieves a 2.9% stockout probability and a 1.1% inventory-full (potential wastage) probability within a 200-step horizon, primarily attends to the age distribution of inventory rather than other features such as day of week or pending orders. Action reachability analysis reveals that the policy employs a diverse replenishment strategy, with most order quantities reached quickly, while several are never selected. Counterfactual analysis shows that replacing medium-large orders with smaller ones leaves both safety probabilities nearly unchanged, indicating that these orders are placed in well-buffered inventory states. This first formal verification and explanation of an RL platelet inventory management policy demonstrates COOL-MC's value for transparent, auditable decision-making in safety-critical healthcare supply chain domains.

Dennis Gross

Inventory ControlReinforcement LearningInterpretability in RLFormal Verification
AskPDF
Mar 2, 2026

Efficient RLVR Training via Weighted Mutual Information Data Selection

Reinforcement learning (RL) plays a central role in improving the reasoning and alignment of large language models, yet its efficiency critically depends on how training data are selected. Existing online selection strategies predominantly rely on difficulty-based heuristics, favouring datapoints with intermediate success rates, implicitly equating difficulty with informativeness and neglecting epistemic uncertainty arising from limited evidence. We introduce InSight, an INformation-guided data SamplInG metHod for RL Training, grounded in a weighted mutual information objective. By modeling data outcomes with Bayesian latent success rates, we show that expected uncertainty reduction decomposes into complementary difficulty- and evidence-dependent components, revealing a fundamental limitation of difficulty-only selection. Leveraging this observation, InSight constructs a stable acquisition score based on the mean belief of datapoints' success rather than noisy sampled outcomes, and naturally extends to multi-rollout settings common in reinforcement learning with verifiable rewards (RLVR). Extensive experiments demonstrate that InSight consistently achieves state-of-the-art performance and improves training efficiency, including a +1.41 average gain on Planning & Mathmatics benchmarks, +1.01 improvement on general reasoning, and up to ~2.2x acceleration, with negligible additional computational overhead.

Xinyu Zhou, Boyu Zhu, Haotian Zhang +2

HKUST(GZ) · UCL · Kuaishou Technology +1

Reinforcement LearningRL for Language Model ReasoningSample-Efficient RLReinforcement Learning with Verifiable Rewards
AskPDF
Mar 2, 2026

SEAR: Sample Efficient Action Chunking Reinforcement Learning

Action chunking improves exploration and accelerates value propagation in long-horizon reinforcement learning, but naively applying off-policy methods to the temporally extended action space at reduced decision frequency offsets these gains, leading to poor sample efficiency. Existing online action chunking methods address these issues through computationally expensive critic-only approaches or by relying on offline data. We introduce SEAR, a sample-efficient off-policy algorithm that enables online reinforcement learning with action chunks by addressing both challenges. To handle high-dimensional action sequences, SEAR employs a causal transformer critic trained with multi-horizon targets that provide a training signal for every prefix of an action chunk, effectively increasing the useful gradients per sample. To restore decision frequency, SEAR operates with a receding horizon and random replanning, ensuring uniform state coverage while combining the fast value propagation of large chunks with the reactivity of small ones. SEAR outperforms state-of-the-art online reinforcement learning methods including SimbaV2 on challenging Metaworld manipulation tasks. Beyond online RL, applying SEAR to the offline-to-online method QC improves its performance on the demanding OGBench cube-triple tasks.

C. F. Maximilian Nagy, Onur Celik, Emiliyan Gospodinov +4

Autonomous Learning Robots, Karlsruhe Institute of Technology · FZI Forschungszentrum Informatik, Karlsruhe

Reinforcement LearningSample-Efficient RLOff-Policy RL
AskPDF
Mar 2, 2026

State-Action Inpainting Diffuser for Continuous Control with Delay

Signal delay poses a fundamental challenge in continuous control and reinforcement learning (RL) by introducing a temporal gap between interaction and perception. Current solutions have largely evolved along two distinct paradigms: model-free approaches which utilize state augmentation to preserve Markovian properties, and model-based methods which focus on inferring latent beliefs via dynamics modeling. In this paper, we bridge these perspectives by introducing State-Action Inpainting Diffuser (SAID), a framework that integrates the inductive bias of dynamics learning with the direct decision-making capability of policy optimization. By formulating the problem as a joint sequence inpainting task, SAID implicitly captures environmental dynamics while directly generating consistent plans, effectively operating at the intersection of model-based and model-free paradigms. Crucially, this generative formulation allows SAID to be seamlessly applied to both online and offline RL. Extensive experiments on delayed continuous control benchmarks demonstrate that SAID achieves state-of-the-art and robust performance. Our study suggests a new methodology to advance the field of RL with delay.

Dongqi Han, Wei Wang, Enze Zhang +1

Microsoft Research Asia · University of Pennsylvania · Shanghai Jiaotong University

Delayed RLReinforcement LearningRL ControlModel-Based RLModel-Free RL
AskPDF
Mar 2, 2026

GAC: Stabilizing Asynchronous RL Training for LLMs via Gradient Alignment Control

Asynchronous execution is essential for scaling reinforcement learning (RL) to modern large model workloads, including large language models and AI agents, but it can fundamentally alter RL optimization behavior. While prior work on asynchronous RL focuses on training throughput and distributional correction, we show that naively applying asynchrony to policy-gradient updates can induce qualitatively different training dynamics and lead to severe training instability. Through systematic empirical and theoretical analysis, we identify a key signature of this instability: asynchronous training exhibits persistently high cosine similarity between consecutive policy gradients, in contrast to the near-orthogonal updates observed under synchronized training. This stale-aligned gradient effect amplifies correlated updates and increases the risk of overshooting and divergence. Motivated by this observation, we propose GRADIENT ALIGNMENT CONTROL(GAC), a simple dynamics-aware stabilization method that regulates asynchronous RL progress along stale-aligned directions via gradient projection. We establish convergence guarantees under bounded staleness and demonstrate empirically that GAC recovers stable, on-policy training dynamics and matches synchronized baselines even at high staleness.

Haofeng Xu, Junwei Su, Yukun Tian +3

School of Computing and Data Science, The University of Hong Kong · Alibaba Group · Southeast University

RL StabilityAsynchronous Advantage Actor-CriticAsynchronous RLReinforcement LearningPolicy Gradient MethodsRL for Language Models
AskPDF
Mar 2, 2026

LLM-assisted Semantic Option Discovery for Facilitating Adaptive Deep Reinforcement Learning

Despite achieving remarkable success in complex tasks, Deep Reinforcement Learning (DRL) is still suffering from critical issues in practical applications, such as low data efficiency, lack of interpretability, and limited cross-environment transferability. However, the learned policy generating actions based on states are sensitive to the environmental changes, struggling to guarantee behavioral safety and compliance. Recent research shows that integrating Large Language Models (LLMs) with symbolic planning is promising in addressing these challenges. Inspired by this, we introduce a novel LLM-driven closed-loop framework, which enables semantic-driven skill reuse and real-time constraint monitoring by mapping natural language instructions into executable rules and semantically annotating automatically created options. The proposed approach utilizes the general knowledge of LLMs to facilitate exploration efficiency and adapt to transferable options for similar environments, and provides inherent interpretability through semantic annotations. To validate the effectiveness of this framework, we conduct experiments on two domains, Office World and Montezuma's Revenge, respectively. The results demonstrate superior performance in data efficiency, constraint compliance, and cross-task transferability.

Chang Yao, Jinghui Qin, Kebing Jin +1

Reinforcement LearningLLM-Guided RLHierarchical RLConstrained RL
AskPDF