Contrastive Branch Policy Optimization (CBPO) is a reinforcement learning method that separates the allocation of a fixed rollout budget from the translation of branch outcomes into token-level credit. It uses generation entropy to screen branch positions, path- and node-level decay to distribute the budget, and Contrastive Branch Value (CBV) to estimate local decision sensitivity without changing reward signs. CBPO partitions trajectories into non-overlapping credit segments, preventing duplicated gradients and enabling fine-grained credit assignment using only outcome rewards.
By Ying Wang, Changlin Qiu, Bang Lin, Linbo Jin, Wen Jiang, Zhe Sun, Jingli Yang
arXiv:2607. 22724v1 Announce Type: cross Abstract: Group-based policy optimization has been increasingly used to train large language model (LLM) agents from sparse outcome rewards by comparing trajectories or steps within a group.
By Kaibing Yang, Guangfeng Cai, Shengtian Yang, Shuo He, Yu Li, Mengyi Liu, Pengwei Chen, Jun Xu, Lei Feng
arXiv:2606. 19236v1 Announce Type: cross Abstract: Reinforcement Learning with Verifiable Rewards algorithms like GRPO have emerged as the dominant post-training paradigm for complex reasoning in LLMs, yet commonly suffer from policy entropy collapse during training.
By Haipeng Luo, Qingfeng Sun, Songli Wu, Can Xu, Wenfeng Deng, Han Hu, Yansong Tang
arXiv:2610.20318v1 Announce Type: cross
Abstract: Reinforcement-learning policies for robot locomotion can fail at deployment when operating conditions shift (e.g., rough or slippery terrain, payload...
By Motasem Alfarra, Pietro Mazzaglia, Markus Peschl, Daniel Dijkman, Christos Louizos
TIGPO (Temporal Instance-Graph Policy Optimization) extends graph-based credit assignment for long-horizon LLM agents by maintaining a persistent transition graph per task across policy updates. It allocates rollout budgets to both new exploration and revisiting past tasks, pairing current rollouts with earlier ones to create cross‑temporal references that stabilize advantage estimation. Experiments on ALFWorld and WebShop show TIGPO consistently outperforms previous group‑based and graph‑based policy optimization methods.
By Jinwei Gan
arXiv:2607. 07178v1 Announce Type: cross Abstract: Recent breakthroughs of Reinforcement Learning (RL) have highlighted its potential for complex agentic Large Language Model (LLM) tasks.
By Zetian Hu, Shunyu Liu, Junjie Zhang, Yongcheng Jing, Ting-En Lin, Yongbin Li, Dacheng Tao
Reinforcement-learning policies for robot locomotion can fail at deployment when operating conditions shift (e.g., rough or slippery terrain, payload changes, sensing degradation), and retraining or m...
arXiv:2607. 25659v1 Announce Type: new Abstract: Rubric-based reinforcement learning enriches language model training by evaluating model outputs against explicit criteria.
By Bo-Wen Zhang, Junwei He, Wen Wang, Song-Lin Lv, Wentao Ma, Rongyi Lin, Shuhan Zhong, Lan-Zhe Guo
The paper introduces Coverage-conditioned Group Policy Optimization (CovGPO), a method that addresses sampling imbalance in group-based policy optimization for long-horizon agentic tasks. CovGPO applies first-visit observation coverage only when all samples in a group receive zero outcome reward, giving higher relative advantages to state-exploring actions without altering the original reward signal. Experiments on ALFWorld and WebShop benchmarks demonstrate that CovGPO consistently outperforms group-based baselines, especially on hard tasks.
By Kaibing Yang, Guangfeng Cai, Shengtian Yang, Shuo He, Yu Li, Mengyi Liu, Pengwei Chen, Jun Xu, Lei Feng
arXiv:2610.02198v1 Announce Type: cross
Abstract: Several state-of-the-art methods for online reinforcement learning in continuous control improve policies using action gradients of a learned critic....
By Sebastian Sanokowski, Alireza Sarmadi, Majid Khadiv
arXiv:2603. 25464v2 Announce Type: replace-cross Abstract: Zero-shot reinforcement learning (RL) algorithms aim to learn a family of policies from a reward-free dataset, and recover optimal policies for any reward function directly at test time.
By Jiajun Hu, Nuria Armengol Urpi, Jin Cheng, Stelian Coros
The paper introduces Flow Matching Policy Optimization with Mirror Descent and Entropy Constraints (FMER), a reinforcement learning framework that uses ODE-based flow matching to overcome limitations of SDE-based diffusion policies. FMER derives an analytic entropy objective to correct density distortion from a tanh transformation, establishes that minimizing an advantage-weighted conditional flow matching loss serves as a simulation-free surrogate for policy mirror descent, and dynamically tunes the mirror descent temperature to enforce a robust trust region. Experiments show FMER outperforms baselines on the sparse-reward FrankaKitchen task while remaining competitive on dense-reward MuJoCo benchmarks.
By Ting Gao, Stavros Orfanoudakis, Nan Lin, Winnie Daamen, Serge Hoogendoorn, Elvin Isufi