The paper introduces GARDEN, a framework that uses large language models to autonomously design a single reward program for cooperative multi‑agent reinforcement learning tasks. By navigating evidence from environment code, training logs, and agent behavior, the reward is tested and finalized without altering the learning algorithm. Experiments on Overcooked and SMAX benchmarks show that the autonomously engineered reward improves task returns by up to 58.2% and win rates by up to 89.9% compared to default rewards, outperforming prior methods.
By Dogan Urgun
The paper introduces Density-Aware Reward Aggregation (DARA), a method that adjusts reward weighting in multi-reward reinforcement learning based on the density of active rewards within rollout batches. By applying an inverse-square-root density correction, DARA gives more weight to less frequently active rewards, enabling faster learning of targeted behaviors. Experiments on tool calling and mathematical reasoning demonstrate that DARA achieves comparable final performance while reducing training steps by up to 26% and 65% respectively.
By Tong Zheng, Skylar Zhai, Zhan Cheng, TianMing Sha, Youling Huang, Shuo Zhou, Shaotong Qi, Jingcheng Liang, Xuwei Ding, Pengcheng Xu
arXiv:2607. 04713v1 Announce Type: cross Abstract: Reinforcement learning holds significant potential for training large language models (LLMs) to handle multi-turn interactive tasks.
By Qiang Liu, Taian Guo, Ruizhi Qiao, Xing Sun
arXiv:2512.13399v3 Announce Type: replace-cross
Abstract: Crafting effective reward signals remains a central challenge in reinforcement learning (RL) for complex reasoning. Existing methods for auto...
By Sitao Cheng, Tianle Li, Xuhan Huang, Xunjian Yin, Difan Zou
The paper introduces Instruct-to-Act, a system that decouples planning and control by combining a vision‑language model (VLM) planner with a world‑model controller. The VLM generates sparse, high‑level text instructions, while the controller executes them at high frequency, trained via relabeling rollouts with synthetic instructions and joint optimization of behavior cloning, reward, and world‑model objectives. Across seven embodied environments—including multi‑agent settings—this approach outperforms controller‑only and direct VLM action methods, maintains fast control, and allows swapping pretrained VLM planners without fine‑tuning, achieving competitive results with strong baselines on most tasks.
By Zineng Tang, Kelsey R. Allen, Sjoerd van Steenkiste, Ishita Dasgupta, Alane Suhr
PlanPO introduces a group planning-aware policy optimization method for multi-turn agentic large language models, addressing the issue of advantage collapse caused by treating all successful trajectories equally. By incorporating coarse-to-fine advantage signals that reflect differences in trajectory and turn lengths, PlanPO encourages agents to learn generalizable planning and generation behaviors. Experiments show a 27.2% average improvement over GRPO on benchmarks such as ALFWorld, WebShop, and SciWorld, with minimal extra training cost.
By Dayang Liang, Liyuan He, Xuan Feng, Shuxin Li, Bo An, Yunlong Liu