The paper introduces a search‑aware reinforcement learning framework for multi‑component query understanding in Roblox game search. It first uses teacher‑student supervised fine‑tuning to create a schema‑compliant policy, then applies reinforcement learning that optimizes each query‑understanding component with component‑specific rewards derived from live search engine interactions. Experiments show that this approach improves per‑component utility and overall search quality, raising NDCG@20 by 8.9 points over the supervised baseline and 3.5 points over a single end‑to‑end reward strategy.
By Nayoung Choi, Shengjian Chen, Xiaokai Wei, Wenzheng Zhang, Daiyao Yi, Rachit Pareek, Vincent Su, Michelle Gong, Jinho D. Choi
The paper introduces a protocol for auditing and composing reinforcement‑learning policies using discrete behavioral rules, defining auditability through six testable predicates such as trace integrity and rule coverage. Experiments show that overlapping rule sets do not guarantee behavioral agreement, and that rule‑based fusion often fails to outperform value‑based composition, highlighting limitations in current description layers. The authors provide an evidence‑bounded audit framework and outline future directions for more robust skill composition.
By Liu Hung Ming
The paper introduces Selective Supervision for Direct-OPD (S$^2$D-OPD), a refinement of Direct On-Policy Distillation that filters out states where the teacher’s policy change is minimal, as measured by the teacher‑reference Jensen‑Shannon divergence. By masking low‑divergence states and keeping only the top 10% of states per response, S$^2$D-OPD improves held‑out accuracy on AIME and HMMT benchmarks across multiple teacher‑student pairs without additional forward passes.
By Yibo Zhao, Zixuan Yang, Yunshi Lan, Xiang Li
The paper presents a reinforcement‑learning approach to schedule link‑level entanglement in quantum networks, using a Markov Decision Process and double deep Q‑networks with message‑passing neural networks. The resulting policies achieve 100% success rates even when the link activation probability is reduced by up to 71% compared to baseline heuristics, and maintain at least 80% success when task placements are hardware‑restricted. The authors also develop metrics to interpret the learned policy and employ a large language model to generate a heuristic that matches the DQN performance, suggesting a scalable method for extracting interpretable strategies in large quantum networks.
By Leon Rode, Sumeet Khatri, Supartha Podder
The paper introduces MI‑SARSA, an on‑policy temporal‑difference algorithm that incorporates mutual‑information regularization to model bounded rationality in reinforcement learning. By penalizing state‑specific deviations from a learned marginal action prior, the algorithm selectively uses state information only when the expected return outweighs the informational cost, yielding a reward‑complexity tradeoff. MI‑SARSA also predicts reaction times, showing that stronger information penalties lead to simpler policies, lower control costs, and faster responses, while regularization mitigates performance loss after environmental shifts at the expense of asymptotic return.
By James Wu, Chris R. Sims
The paper proposes using lightweight, calibrated System One decision models—specifically JEV and Laya—to improve autonomous penetration-testing harnesses that rely on large language models (LLMs). It defines four key decision points (finding adjudication, severity recalibration, agent pruning, and confirmation loops) and presents a NeuroSploit case study showing differences in severity distribution, runtime, and grading when using TypeSafe System One. The authors review existing System One specifications, discuss various RL-based training approaches, and introduce Rave, a domain‑adapted model with a proposed training and evaluation framework.
By Joas Antonio dos Santos Barbosa
The paper introduces AdvRole, an adversarial closed‑loop curriculum for training role‑playing agents with large language models. It alternates between an Actor that learns to role‑play and a Rewriter that edits character profiles and dialogue contexts into hard scenarios, using a performance‑gap reward to target the Actor’s weaknesses. Experiments on English, Chinese, and a new multilingual benchmark demonstrate that AdvRole consistently outperforms baseline methods.
By Zheng Zhang, Liu Liu, Qi Chai, Deheng Ye, Peilin Zhao, Mao Zheng, Hao Wang
The paper introduces Reinforcement Learning with Verifiable Rewards (RLVR) applied to small search agents, specifically training a Qwen3.5-0.8B model with Group Relative Policy Optimization and an interleaved Wikipedia-search tool on the MuSiQue dataset. Experiments varying reward shapes across three seeds show that RLVR can achieve a 3.8‑fold improvement over an untrained baseline, with the best run reaching a 0.352 average exact match. The study finds that the sparse exact‑match reward, standard in larger models, performs poorly for small models, indicating that reward design must be tailored rather than scaled down from large‑model recipes.
By Gaurisankar Jayadas, Aske Plaat, \'Alvaro Serra-G\'omez, Sandheep P
The paper introduces GRAFT, a Graph-based Faithful sTep-level credit-assignment framework that constructs a trajectory graph from rollout trajectories, recovers node state-values via Bellman iteration, and assigns step-level advantages based on node value differences. It also proposes Graph GAE to further reduce state-value estimation bias. Experiments on multi-turn agentic benchmarks demonstrate consistent improvements over GRPO and other recent agentic RL algorithms.
By Xincheng Yao, Haobo Fu, Weiming Liu, Chongyang Zhang
SLCA-GRPO addresses cross‑segment credit misattribution in tool‑calling reinforcement learning by introducing Segment‑Locked Credit Assignment (SLCA), which separates advantage estimation for tool‑invocation tokens and natural‑language summary tokens. The method leverages a Schema‑Guided LLM Simulator (SGLS) for scalable training and Hierarchical Rewards (HierR) to route execution and preference advantages appropriately. Experiments on a 7B backbone show that SLCA‑GRPO outperforms baseline methods, improving in‑domain accuracy by 2.53 pp, the Berkeley Function‑Calling Leaderboard by 1.36 pp, and $ au^2$‑Bench by 9.15 pp while reducing tool redundancy and costs.
By Yan Zhan, Shaobo Liu, Qiunan Liu, Yuanjun Shi, Siqi Xu, WeiYi Hou, Xiang Xu, Zekang Li, Weizhou Pan, Jiahong Yan
The paper proposes a reinforcement‑learning framework, termed Fourth Degree Learning, that automatically selects the most suitable classification algorithm for disease prediction tasks. Using this method, the authors report a significant improvement in predictive accuracy for Primary Biliary Cirrhosis, raising it from 63% to 98%. The approach relies on a scoring mechanism called square learning (SL) to guide the algorithm selection process.
By Hasan Samadbin, Arman Daliri
The paper introduces Jev, a reinforcement‑learning‑trained model that provides calibrated probability answers to typed questions about a single input in one call. Jev is evaluated on RLCDAlignBench, a benchmark covering ten alignment failures across 44 tests and five target models, achieving a median AUROC of 0.886 zero‑shot and outperforming supervised baselines on most tasks. The study shows that question wording has little impact, while contextual fields that encode labels are more influential, and that Jev matches human‑label agreement while being 63× cheaper than LLM‑judge scorers.
By Ruoqi Guo, Yi Liu, Gelei Deng, Yuekang Li, Lida Zhao, Yutao Wu, Simin Chen, Ying Zhang, Leo Yu Zhang
The paper introduces iCoder-27B, a 27‑billion‑parameter model for RTL design and GPU kernel optimization that is developed through a recursive AI‑led process with minimal human input. Human experts provide high‑level objectives and reusable research skills, while the agent autonomously selects experiments, diagnoses outcomes, and refines training strategies, coordinating SFT, self‑distillation, and reinforcement learning. iCoder outperforms GPT‑5.5 and Claude‑Opus‑4.8 on several benchmarks, demonstrating the feasibility of building frontier‑competitive models with largely automated development.
By Cheng Yang, Jiayang Lyu, Shangyuan Liu, Guibin Zhang, Jiong Lin, Xinlei Yu, Junchi Yan, Shuicheng Yan, Weinan E, Linfeng Zhang, Linfeng Zhang, Qibing Ren
Qwen‑Planner‑Agent is a closed‑loop AI‑for‑AI framework that enables large language models to act as both developers and participants in building advanced AI systems. The framework integrates data production, model training, and deployment through a shared action‑feedback‑verification contract, employing AI‑for‑Data, AI‑for‑Training, and AI‑driven model‑harness co‑evolution. It achieves top performance on MobilePA‑Bench by improving tool use, memory, skills, and sub‑agent coordination, while also showing gains on non‑mobile benchmarks.
By Tingyu Qu, Weigao Sun, Yuecheng Liu, Yucheng Zhao, Yi Zhu, Yifeng Ding, Qiyi Wang, Sihan Cao, Pengkun Jiao, Hanlei Xie, Xiongwei Wu, Qichao Wang, Haodong Zhang, Jiajun Liu, Yuhao Wang, Yuqing Xie, Junpeng Zhao, Long Chen, Ming Ma, Sihan Yang, Ziwang Zhao, Yanhao Jia, Liangquan Gong, Feida Zhu, Yiran Zhong, Steven Hoi
The paper introduces Self‑Play Pretraining with Zero Data, a proof‑of‑concept method that lets a model generate its own training data by searching over all computable processes using a universal Turing machine. Two models— a generator that proposes byte‑sequence programs and a learner that predicts those sequences—train together, with the generator rewarded for producing data at the learner’s frontier, creating an adaptive curriculum. Experiments show that zero‑shot performance on natural datasets scales predictably with compute, and the models exhibit in‑context learning and discover mathematical sequences during training.
By Aditya Cowsik, Kfir Dolev, Michael Y. Li, G. Bruno De Luca, Nourya Cohen, Noah D. Goodman, Yoav Levine
The paper presents Aftab, a new architecture for replay‑free parallelized Q‑learning that systematically explores visual encoder designs, multiplicative feature interactions, and value‑estimation strategies. Through a three‑phase study on Atari‑57, the authors compare eight convolutional encoders, integrate Hadamax‑style interactions, and evaluate categorical‑dueling, ensemble‑dueling, and combined configurations, ultimately achieving a higher human‑normalized score than the baseline PQN. Aftab is also evaluated on Procgen Hard, showing improved terminal IQM and learning‑curve area, and the full framework is released as open source.
By Taha Shieenavaz, Shabnam Zareshahraki, Loris Nanni
AV‑GRPO introduces a modality‑anchored diffusion reinforcement learning framework for joint audio‑video generation, addressing limitations in fidelity, text‑modality alignment, and cross‑modal synchronization. It decouples learning signals through modality‑anchored rollouts, employs trajectory‑locked frozen‑tower optimization to reduce computational cost, and adapts objectives to each modality’s dynamics. The accompanying 5DAV dataset provides difficulty‑controllable, decoupled training samples, and experiments on JavisBench and VABench show AV‑GRPO surpasses LTX‑2.3 in generation quality, semantic alignment, and synchronization.
By Zhiyu Xu, Weilong Yan, Yufei Shi, Shiyang Li, Yihao Liu, Kin-Man Lam, Yuewen Cao
The paper introduces Predictive Action Chunk Learning (PACL), a method for improving robot manipulation policies using mixed-quality deployment experience. PACL first trains a predictive chunk-level critic to evaluate temporally extended action sequences, then uses the critic’s quality estimates to guide a diffusion actor that learns from both successful and failed rollouts. Experiments on simulated and real robots demonstrate that PACL consistently enhances pretrained policies and outperforms strong imitation learning and offline reinforcement learning baselines.
By Yangang Ren, Yujie Yan, Zirui Li, Jiaming Guo, Di Zeng, Ji Tao, Lan Yu, Xuesong Tian, Chen Lv
The paper introduces a vector Bellman theory for multichain robust average‑reward Markov decision processes, addressing the state‑dependent optimal long‑run rewards that arise under uncertainty. It develops a gain‑first, bias‑second optimization principle for finite models with compact, post‑action $(s,a)$‑rectangular ambiguity, yielding a coupled vector gain‑bias system and stationary saddle strategies from all initial states. The authors also characterize solvability conditions, provide certificates for asymptotically affine trajectories of the robust Bellman operator, and design a robust approximately shifted Halpern planning algorithm that converges to the optimal gain vector and produces average‑optimal greedy controllers.
By Yue Wang, George Atia
The paper proposes a method for selectively querying language‑model advice in reinforcement learning by predicting the value of potential responses and only querying when the expected benefit outweighs the cost. It introduces a certified, response‑contingent metareasoning framework that guarantees near‑optimal advice usage under certain assumptions, and demonstrates that a calibrated controller with Qwen2.5 advisors can improve task performance while drastically reducing the number of advice calls on the BabyAI benchmark.
By Ibne Farabi Shihab, Md Najmus Swaqeeb, Abu Sa-Adat Mohamed Moon-Im Al Ahsan