ScholarPulse 日报 2026-10-02

2026-10-02 学术简报：2 篇。InterEvolve通过测试时进化奖励程序，使人形机器人无需重训练即可重用现有技能解决新的loco-manipulation任务，并从自身尝试中改进并保留所学。

今日速览

序号	标题	来源	日期	主题	推荐等级
1	InterEvolve: Test-Time Evolution of Reward Programs for Humanoid Loco-Manipulation	arXiv	2026-10-01	AI-Agent	高
2	From Knowledge Access to Source Learning: Developing Source-Specific Competence	arXiv	2026-10-01	RAG	高

重点论文与技术动态

1. InterEvolve: Test-Time Evolution of Reward Programs for Humanoid Loco-Manipulation

来源：arXiv
日期：2026-10-01
作者/机构：Zhuo Lin, Sirui Xu, Liuyu Bian, Yu-Xiong Wang, Liang-Yan Gui
主题标签：AI-Agent, arXiv
推荐等级：高
分类：cs.RO, cs.CV, cs.GR

一句话结论

InterEvolve通过测试时进化奖励程序，使人形机器人无需重训练即可重用现有技能解决新的loco-manipulation任务，并从自身尝试中改进并保留所学。

核心内容

广泛控制器已持有新任务所需能力，通过规划与控制间表达接触丰富多阶段交互且可执行可测量的接口访问。
对象感知前向-后向行为基础模型将新奖励转化为测试时loco-manipulation行为。
任务指定为奖励程序，LLM代理修订结构利用执行反馈和技能库，数值优化器调整常数，迭代改进。

方法与数据

方法为对象感知FB模型结合LLM修订奖励程序和数值优化，候选在并行仿真场景验证。
实验表明进化程序优于人类设计奖励，支持多样任务、复杂场景和长时程组合，并在物理Unitree G1上自主运行。摘要未明确具体数据。

价值判断

值得关注：无需重训练释放FB模型潜力，有时产生新颖策略。
可复用点：奖励程序作为接口，LLM结合技能库进化。
局限/待核查：物理实验细节有限，摘要无定量指标。

摘要

We study test-time evolution for humanoid loco-manipulation: solving tasks that a controller was never trained for by repurposing its existing skills, improving from its own attempts, and retaining what it learns, without retraining. Our key insight is that a broad controller already holds much of the competence a new task needs, and that this competence becomes accessible through an interface between planning and control that is expressive enough to specify contact-rich, multi-stage interactions, yet executable and measurable enough that execution feedback can guide planning from experience. InterEvolve realizes this interface with two components. First, we develop an object-aware forward-backward (FB) behavioral foundation model, whose object residuals on a frozen body prior turn a new reward about the body or objects into loco-manipulation behavior at test time. Second, we specify tasks as reward programs: staged rewards with completion conditions and tunable constants. A large language model (LLM) agent revises the program structure in context, drawing on execution feedback and a skill library of verified programs, while a numerical optimizer tunes its constants. With every candidate verified across parallel simulation scenarios, the program explores new ways to induce, repurpose, and compose the controller's existing motor competence for the task at hand, and thus improves over iterations. Experiments show that human-designed rewards leave much of the FB model's loco-manipulation competence untapped, whereas the programs InterEvolve evolves release it, sometimes through novel strategies. It further produces behaviors for diverse tasks, complex scenes, and long-horizon compositions in simulation, and evolved skills run autonomously on a physical Unitree G1 from egocentric onboard perception.

2. From Knowledge Access to Source Learning: Developing Source-Specific Competence

来源：arXiv
日期：2026-10-01
作者/机构：Lucheng Fu, Kejing Xia, Yiyang Wang, Yiqiao Jin, Jinjin He, Xiyuan Yang
主题标签：RAG, arXiv
推荐等级：高
分类：cs.CL, cs.AI, cs.LG

一句话结论

SourceLearn通过持续来源模型，将智能体对权威来源的重复使用从反复访问转为可复用来源特定能力学习，多数设置优于检索与记忆基线。

核心内容

现有方法优化访问组织或保存交互记忆，未把同一来源反复使用视为加深理解的机会。
来源学习以持续来源模型刻画知识的结构、解释与应用，形成可复用能力。
SourceLearn结合自导向重访未充分理解处与任务引导补缺口；信号定范围，更新从权威源重建。

方法与数据

两类机制：自导向与任务引导来源学习，更新从权威来源重建。
五基准、三后端；15设置中13最优，较Hybrid RAG最高+22.6，整体优于静态表征与经验记忆基线。

价值判断

值得关注：从知识访问转向来源学习，权威重建保忠实。
可复用点：双学习信号与“信号定范围、权威源重建内容”。
局限/待核查：摘要未明确基准名称、来源类型与计算开销。

摘要

Large language model (LLM) agents increasingly rely on persistent external sources to solve sequences of knowledge-intensive tasks. Existing methods improve how source content is accessed and organized, while agent-memory systems preserve reusable knowledge from prior interactions, but repeated use of the same source is still largely treated as repeated access rather than an opportunity to progressively improve understanding of that source. We study source learning: developing reusable source-specific competence over a persistent authoritative source. We represent this competence with a persistent source model that captures reusable understanding of the source, including how its knowledge is structured, interpreted, and applied. To construct and progressively refine such models, we propose SourceLearn, which combines two complementary learning mechanisms. Self-Directed Source Learning identifies what remains incompletely understood and adaptively revisits the source, while Task-Guided Source Learning uses downstream experience to reveal local representational gaps and recurring needs in how source knowledge should be organized. In both cases, learning signals determine what should be reconsidered, while persistent updates are reconstructed from the authoritative source. Across five benchmarks and three LLM backends, SourceLearn achieves the best performance in 13 of 15 settings, with gains of up to 22.6 points over Hybrid RAG and substantial overall improvements over static source representations and experience-based memory baselines.