一种自我纯净轨迹方法，大幅提升4B-7B小模型的Agentic强化学习效果，专治执行失败噪声！
arXiv:2601.15141v2 Announce Type: replace Abstract: Agentic Reinforcement Learning (RL) has empowered Large Language Models (LLMs) to utilize tools li…
用最优性先验引导神经网络，高效求解月球着陆能量最优轨迹，为航天制导带来新思路。
arXiv:2607.02741v1 Announce Type: cross Abstract: This paper develops an Optimality-Informed Neural Network (OINN) approach for the energy-optimal, fr…
本地调试AI代理轨迹的RLM优化器，GitHub开源，支持生产数据训练。
Hi all! We built HALO, an open-source local tool for debugging and improving agent harnesses from AI agent traces. It ingests Langfuse, Arize/OpenInfe…
用LLM优化异构无人机网络中的保密预编码与轨迹，RSMA多址技术破解安全通信难题
arXiv:2507.17188v2 Announce Type: replace-cross Abstract: This paper investigates secure communications in rate-splitting multiple access (RSMA) enabl…
一篇统一SFT、DAgger、离线RL和OPD视角的LLM蒸馏论文，解耦KL与轨迹，为模型优化提供新理论框架。
arXiv:2605.16826v1 Announce Type: new Abstract: Knowledge distillation is central to LLM post-training, yet its design space remains poorly understood…