In reinforcement learning, the softmax parametrization is the standard approach for policies over discrete action spaces. However, it fails to capture the order relationship between actions. Motivated by a real-world industrial problem, we propose a novel policy parametrization based on ordinal regression models adapted to the reinforcement learning setting. Our approach addresses practical challenges, and numerical experiments demonstrate its effectiveness in real applications and in continuous action tasks, where discretizing the action space and applying the ordinal policy yields competitive performance.

--The social metaverse is a growing digital ecosystem that blends virtual and physical worlds. It allows users to interact socially, work, shop, and enjoy entertainment. However, privacy remains a major challenge, as immersive interactions require continuous collection of biometric and behavioral data. At the same time, ensuring high-quality, low-latency streaming is difficult due to the demands of real-time interaction, immer-sive rendering, and bandwidth optimization. T o address these issues, we propose ASMS (Adaptive Social Metaverse Streaming), a novel streaming system based on Federated Multi-Agent Proximal Policy Optimization (F-MAPPO). ASMS leverages F-MAPPO, which integrates federated learning (FL) and deep reinforcement learning (DRL) to dynamically adjust streaming bit rates while preserving user privacy. Experimental results show that ASMS improves user experience by at least 14% compared to existing streaming methods across various network conditions. Therefore, ASMS enhances the social metaverse experience by providing seamless and immersive streaming, even in dynamic and resource-constrained networks, while ensuring that sensitive user data remains on local devices. Index T erms --Social metaverse, adaptive bit rate streaming, Multi-agent reinforcement learning, federated learning, extended reality. The metaverse is seen as the next evolution of the Internet, offering a seamless digital space where users can meet, socialize, play games, and collaborate in immersive 3D environments [1]. As adoption grows, it has gained significant global attention. Gartner predicts that by 2026, 25% of people will spend at least an hour per day in metaverse environments [2].

Resource constraints can fundamentally change both learning and decision-making. We explore how memory constraints influence an agent's performance when navigating unknown environments using standard reinforcement learning algorithms. Specifically, memory-constrained agents face a dilemma: how much of their limited memory should be allocated to each of the agent's internal processes, such as estimating a world model, as opposed to forming a plan using that model? We study this dilemma in MCTS- and DQN-based algorithms and examine how different allocations of memory impact performance in episodic and continual learning settings.

Recent advances in large reasoning models (LRMs) demonstrate that sophisticated behaviors such as multi-step reasoning and self-reflection can emerge via reinforcement learning with verifiable rewards~(\textit{RLVR}). However, existing \textit{RLVR} approaches are inherently ``on-policy'', limiting learning to a model's own outputs and failing to acquire reasoning abilities beyond its initial capabilities. To address this issue, we introduce \textbf{LUFFY} (\textbf{L}earning to reason \textbf{U}nder o\textbf{FF}-polic\textbf{Y} guidance), a framework that augments \textit{RLVR} with off-policy reasoning traces. LUFFY dynamically balances imitation and exploration by combining off-policy demonstrations with on-policy rollouts during training. Specifically, LUFFY combines the Mixed-Policy GRPO framework, which has a theoretically guaranteed convergence rate, alongside policy shaping via regularized importance sampling to avoid superficial and rigid imitation during mixed-policy training. Compared with previous RLVR methods, LUFFY achieves an over \textbf{+6.4} average gain across six math benchmarks and an advantage of over \textbf{+6.2} points in out-of-distribution tasks. Most significantly, we show that LUFFY successfully trains weak models in scenarios where on-policy RLVR completely fails. These results provide compelling evidence that LUFFY transcends the fundamental limitations of on-policy RLVR and demonstrates the great potential of utilizing off-policy guidance in RLVR.

The increasingly challenging task of maintaining power grid security requires innovative solutions. Novel approaches using reinforcement learning (RL) agents have been proposed to help grid operators navigate the massive decision space and nonlinear behavior of these complex networks. However, applying RL to power grid security assessment, specifically for combinatorially troublesome contingency analysis problems, has proven difficult to scale. The integration of quantum computing into these RL frameworks helps scale by improving computational efficiency and boosting agent proficiency by leveraging quantum advantages in action exploration and model-based interdependence. To demonstrate a proof-of-concept use of quantum computing for RL agent training and simulation, we propose a hybrid agent that runs on quantum hardware using IBM's Qiskit Runtime. We also provide detailed insight into the construction of parameterized quantum circuits (PQCs) for generating relevant quantum output. This agent's proficiency at maintaining grid stability is demonstrated relative to a benchmark model without quantum enhancement using N-k contingency analysis. Additionally, we offer a comparative assessment of the training procedures for RL models integrated with a quantum backend.

To apply reinforcement learning to safety-critical applications, we ought to provide safety guarantees during both policy training and deployment. In this work, we present theoretical results that place a bound on the probability of violating a safety property for a new task-specific policy in a model-free, episodic setting. This bound, based on a maximum policy ratio computed with respect to a 'safe' base policy, can also be applied to temporally-extended properties (beyond safety) and to robust control problems. To utilize these results, we introduce SPoRt, which provides a data-driven method for computing this bound for the base policy using the scenario approach, and includes Projected PPO, a new projection-based approach for training the task-specific policy while maintaining a user-specified bound on property violation. SPoRt thus enables users to trade off safety guarantees against task-specific performance. Complementing our theoretical results, we present experimental results demonstrating this trade-off and comparing the theoretical bound to posterior bounds derived from empirical violation rates.

Inverse design of photonic integrated circuits (PICs) has traditionally relied on gradient-based optimization. However, this approach is prone to end up in local minima, which results in suboptimal design functionality. As interest in PICs increases due to their potential for addressing modern hardware demands through optical computing, more adaptive optimization algorithms are needed. We present a reinforcement learning (RL) environment as well as multi-agent RL algorithms for the design of PICs. By discretizing the design space into a grid, we formulate the design task as an optimization problem with thousands of binary variables. We consider multiple two-and three-dimensional design tasks that represent PIC components for an optical computing system. By decomposing the design space into thousands of individual agents, our algorithms are able to optimize designs with only a few thousand environment samples. They outperform previous state-of-the-art gradient-based optimization in both two-and three-dimensional design tasks. Our work may also serve as a benchmark for further exploration of sample-efficient RL for inverse design in photonics.

We present the Multi-Agent Transformer World Model (MATWM), a novel transformer-based world model designed for multi-agent reinforcement learning in both vector- and image-based environments. MATWM combines a decentralized imagination framework with a semi-centralized critic and a teammate prediction module, enabling agents to model and anticipate the behavior of others under partial observability. To address non-stationarity, we incorporate a prioritized replay mechanism that trains the world model on recent experiences, allowing it to adapt to agents' evolving policies. We evaluated MATWM on a broad suite of benchmarks, including the StarCraft Multi-Agent Challenge, PettingZoo, and MeltingPot. MATWM achieves state-of-the-art performance, outperforming both model-free and prior world model approaches, while demonstrating strong sample efficiency, achieving near-optimal performance in as few as 50K environment interactions. Ablation studies confirm the impact of each component, with substantial gains in coordination-heavy tasks.

Large vision-language models (VLMs) for autonomous driving (AD) are evolving beyond perception and cognition tasks toward motion planning. However, we identify two critical challenges in this direction: (1) VLMs tend to learn shortcuts by relying heavily on history input information, achieving seemingly strong planning results without genuinely understanding the visual inputs; and (2) the chain-ofthought (COT) reasoning processes are always misaligned with the motion planning outcomes, and how to effectively leverage the complex reasoning capability to enhance planning remains largely underexplored. In this paper, we start from a small-scale domain-specific VLM and propose Drive-R1 designed to bridges the scenario reasoning and motion planning for AD. Drive-R1 first undergoes the supervised finetuning on a elaborate dataset containing both long and short COT data. Drive-R1 is encouraged to reason step-by-step from visual input to final planning decisions. Subsequently, Drive-R1 is trained within a reinforcement learning framework that incentivizes the discovery of reasoning paths that are more informative for planning, guided by rewards based on predicted trajectories and meta actions. Experimental evaluations on the nuScenes and DriveLM-nuScenes benchmarks demonstrate that Drive-R1 achieves superior performance compared to existing state-of-the-art VLMs. We believe that Drive-R1 presents a promising direction for bridging reasoning and planning in AD, offering methodological insights for future research and applications.

--Multiple quadrotor unmanned aerial vehicle (UA V) systems have garnered widespread research interest and fostered tremendous interesting applications, especially in multi-constrained pursuit-evasion games (MC-PEG). The Cooperative Evasion and Formation Coverage (CEFC) task, where the UA V swarm aims to maximize formation coverage across multiple target zones while collaboratively evading predators, belongs to one of the most challenging issues in MC-PEG, especially under communication-limited constraints. This multifaceted problem, which intertwines responses to obstacles, adversaries, target zones, and formation dynamics, brings up significant high-dimensional complications in locating a solution. In this paper, we propose a novel two-level framework (i.e., Consensus Inference-based Hierarchical Reinforcement Learning (CI-HRL)), which delegates target localization to a high-level policy, while adopting a low-level policy to manage obstacle avoidance, navigation, and formation. Specifically, in the high-level policy, we develop a novel multi-agent reinforcement learning module, Consensus-oriented Multi-Agent Communication (ConsMAC), to enable agents to perceive global information and establish consensus from local states by effectively aggregating neighbor messages. Meanwhile, we leverage an Alternative Training-based Multi-agent proximal policy optimization (A T -M) and policy distillation to accomplish the low-level control. The experimental results, including the high-fidelity software-in-the-loop (SITL) simulations, validate that CI-HRL provides a superior solution with enhanced swarm's collaborative evasion and task completion capabilities. Nowadays, quadrotor Unmanned Aerial V ehicles (UA Vs) have demonstrated great potential in costly or human-unfriendly tasks (e.g., disaster response [1]), due to their agility, cost-effectiveness, and compact size. Nevertheless, the UA V swarm is likely to be exposed to an adversarial environment, where a hostile factor or agent might attack the affiliated members, and must respond promptly to boost the survival opportunity. Y uming Xiang and Sizhao Li and Rongpeng Li are with the College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou 310027, China (email: {xiangym1999; liszh5; lirongpeng }@zju.edu.cn).