PulseAugur
中
实时 21:46:49
English(EN) CopilotKit has released CLiF, a continual learning platform that turns agent traces and user actions into skills, evaluations and fine-tuning data. The platform

新框架增强 AI 代理的持续学习和自我改进 · 跟踪 10 个来源

研究人员正在开发新的方法来增强 AI 代理的持续学习能力,使其能够在不忘记先前任务的情况下适应新任务。一种方法 TSRL,通过选择性地将监督应用于困难状态,将监督微调与强化学习相结合。另一种方法,激活条件选择性知识控制,使用神经元级别的梯度操纵来保护历史知识,同时适应新应用。此外,正在创建 EmbodiedCLBench 等新基准来评估具身代理的持续学习能力,并提出了 ReHarness 和 VeriFine 等框架来增强代理的自我改进和验证过程。 AI

影响 这些在持续学习和自我改进方面的进展对于开发能够长期运行和执行复杂任务的更具适应性和鲁棒性的 AI 代理至关重要。

排序理由 多篇研究论文介绍了 AI 代理持续学习的新方法和新基准。

在 Mastodon — mastodon.social 阅读 →

AI 生成摘要 · Google Gemini · 来自 19 个来源。 我们如何撰写摘要 →

新框架增强 AI 代理的持续学习和自我改进 · 跟踪 10 个来源

本文如何被排名

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
多篇研究论文介绍了 AI 代理持续学习的新方法和新基准。
Source corroboration
19 independent sources
Strong cross-source corroboration — multiple independent publishers covered this within the clustering window.
Topics
paper, model release
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
47 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.
Coverage growth since scoring
+3 source(s) since last score
New sources have picked up this story since our last re-score. Score will update on the next scoring pass.

完整方法见我们的编辑标准。

报道来源 [19]

  1. arXiv cs.AI TIER_1 English(EN) · Yixu Huang, Jiaqi W. Ma ·

    面向持续学习智能体的可靠知识获取

    arXiv:2610.21739v1 Announce Type: cross Abstract: As agents powered by foundation models move into open-ended environments, continual learning from experience becomes essential. For agents built on models with fixed parameters, this learning requires reusing prior task experience…

  2. arXiv cs.AI TIER_1 English(EN) · Zirui Shang, Xin Shu, Yang Liu, Zhi Gao, Xinxiao Wu, Lifeng Fan ·

    面向应用流的持续GUI智能体学习的选择性知识控制

    arXiv:2609.06530v1 Announce Type: cross Abstract: Continual learning is a crucial capability for Graphical User Interface (GUI) agents to adapt to evolving applications while retaining knowledge acquired from previous applications. Such application streams pose a challenging know…

  3. arXiv cs.AI TIER_1 English(EN) · Yuqian Fu, Kaiwen Jiang, Zijun Song, Jiawei Wang, Yuanheng Zhu, Dongbin Zhao ·

    TSRL:基于触发器的监督强化学习用于持续智能体训练

    arXiv:2610.17726v1 Announce Type: cross Abstract: Large language model (LLM) agents excel at complex, multi-step tasks but suffer from catastrophic forgetting during continual learning--acquiring new skills often degrades previously mastered capabilities. Existing post-training p…

  4. arXiv cs.AI TIER_1 English(EN) · Yuxin Xiao, Zhiyun Lu, Jianyu Wang, Yangtian Zhang, Yang Song, Qifan Wang, Bailin Wang ·

    用于智能体持续学习的循环式约束器

    arXiv:2610.12668v1 Announce Type: cross Abstract: Language model agents increasingly operate over long interaction streams, requiring online adaptation to distribution shifts through continual learning. However, existing agent harnesses typically either evolve their states only t…

  5. arXiv cs.AI TIER_1 English(EN) · Jie Huang, Yanan Chen, Ruixun Liu, Kaichen He, Xinyi He, Rui Huang, Zilong Zheng, Zhenliang Zhang, Yiwu Zhong ·

    EmbodiedCLBench:评估自进化具身智能体的持续学习能力

    arXiv:2610.13021v1 Announce Type: new Abstract: Continual learning requires models to accumulate past experience and transfer it to new tasks. Despite substantial research on continual learning, existing work remains largely confined to non-embodied settings, without extending to…

  6. arXiv cs.AI TIER_1 English(EN) · Bohan Zhou, Xingbei Chen, Emily Huang, Weilin Ruan, Haojian Huang, Yehang Zhang, Zexi Li, Wenqian Li, Qize Yu, Zetian Song, Leyi Wu, Jinghao Li, Mingxuan Song, Xinrun Xu, Zongyang Qiu, Yangkai Wei, Tianyi Zhang, Kaiwen Zhou, Yinchuan Li, James Cheng ·

    RoboAware:从反事实结果中学习协调具身技能

    arXiv:2610.11480v1 Announce Type: cross Abstract: Embodied coding agents can combine modular robot skills with frozen end-to-end policies, yet effective composition requires anticipating which policy family will succeed in the current physical state. We present RoboAware, which b…

  7. Hugging Face Daily Papers TIER_1 English(EN) ·

    EmbodiedCLBench:评估自进化具身智能体的持续学习能力

    Continual learning requires models to accumulate past experience and transfer it to new tasks. Despite substantial research on continual learning, existing work remains largely confined to non-embodied settings, without extending to embodied environments. To address this gap, we …

  8. Hugging Face Daily Papers TIER_1 English(EN) ·

    用于智能体持续学习的循环式约束器

    Language model agents increasingly operate over long interaction streams, requiring online adaptation to distribution shifts through continual learning. However, existing agent harnesses typically either evolve their states only through forward information flow or rely on additio…

  9. Hugging Face Daily Papers TIER_1 English(EN) ·

    具身图灵机:用于机器人递归自我的状态化代码

    Most robot policies keep a model in the control loop: a VLA maps observations to actions, and an Agent Harness, such as Agent-as-Policy or Harness VLA queries a VLM for decision making at run time. We propose a different view: the embodied world is an Embodied Turing Machine, who…

  10. arXiv cs.AI TIER_1 English(EN) · Zewei Zhou, Rachel Luo, Yulong Cao, Chaowei Xiao, Chensheng Peng, Boyi Li, Thomas Tian, Zheng Lian, Yan Wang, Jiaqi Ma, Boris Ivanovic, Marco Pavone, Wenhao Ding ·

    VeriFine: 扩展自监督学习中的验证以实现自我改进

    arXiv:2610.08761v1 Announce Type: new Abstract: Self-improving policies continually expose new failure patterns, changing what their judges must be able to verify. However, current fixed judges constrain both optimization feedback and the discovery of useful training examples, li…

  11. Hugging Face Daily Papers TIER_1 English(EN) ·

    VeriFine:为具身推理的自我改进扩展验证

    Self-improving policies continually expose new failure patterns, changing what their judges must be able to verify. However, current fixed judges constrain both optimization feedback and the discovery of useful training examples, limiting further self-improvement. This challenge …

  12. Hugging Face Daily Papers TIER_1 English(EN) ·

    EmbodiedSmith:通过模拟中的递归自我改进飞轮扩展具身数据

    Scaling robotic foundation models requires diverse training data and reliable evaluation environments. Simulation offers a scalable solution, yet existing generation pipelines remain constrained by predefined assets and skills, a disconnect between scene generation and task gener…

  13. arXiv cs.CV TIER_1 English(EN) · Kaipeng Fang, Hangxu Liu, Chaojian Shi, Yunsong Zhou, Hongzi Zhu, Minyi Guo, Jingkuan Song, Lianli Gao ·

    面向带修正生成重放的世界动作模型的无数据持续学习

    arXiv:2610.18038v2 Announce Type: replace-cross Abstract: World action models (WAMs) built on pretrained video generation models adapt quickly to new robotic tasks, but they forget previously learned tasks under sequential fine-tuning. Experience replay (ER) mitigates such forget…

  14. arXiv cs.CV TIER_1 English(EN) · Kairui Hu, Siyuan Hu, Fangzhou Hong, Zhaoxi Chen, Ziwei Liu ·

    具身图灵机:用于机器人递归自我的状态化代码

    arXiv:2610.12369v1 Announce Type: cross Abstract: Most robot policies keep a model in the control loop: a VLA maps observations to actions, and an Agent Harness, such as Agent-as-Policy or Harness VLA queries a VLM for decision making at run time. We propose a different view: the…

  15. arXiv cs.CV TIER_1 English(EN) · Yikai Qin, Yifei Deng, Mingjian Liang, Wenxuan Song, Zepeng Lin, Zhiyi Jiang, Jiajun Fu, Qiao Sun, Huashuo Lei, Xicheng Gong, Jiayi Chen, Han Zhao, Shuanghao Bai, Pengxiang Ding, Pengwei Wang, Haoang Li ·

    EmbodiedSmith:通过模拟中的递归自我改进飞轮扩展具身数据

    arXiv:2610.07969v1 Announce Type: new Abstract: Scaling robotic foundation models requires diverse training data and reliable evaluation environments. Simulation offers a scalable solution, yet existing generation pipelines remain constrained by predefined assets and skills, a di…

  16. MarkTechPost TIER_1 English(EN) · Asif Razzaq ·

    CopilotKit 发布 CLiF:一个将代理轨迹和用户操作转化为技能、评估和微调数据的持续学习平台

    <p>Introduction Teams shipping AI agents at scale keep running into the same pattern. Even agents built on top-tier models like Fable and Astra struggle to carry out complex, end-to-end tasks inside a specific product consistently enough to deploy. So teams grind out evals for th…

  17. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] ·

    📰 CopilotKit 发布 CLiF:一个将代理痕迹和用户操作转化为技能、评估和微调数据的持续学习平台 介绍团队

    📰 CopilotKit Releases CLiF: A Continual Learning Platform That Turns Agent Traces and User Actions into Skills, Evals, and Fine-Tuning Data Introduction Teams shipping AI agents at scale keep running into the same pattern. Even agents built on top-tier models like Fable and Astra…

  18. Mastodon — mastodon.social TIER_1 English(EN) · [email protected] ·

    📰 CopilotKit 发布 CLiF:一个将代理轨迹和用户操作转化为技能、评估和微调数据的持续学习平台 介绍团队

    📰 CopilotKit Releases CLiF: A Continual Learning Platform That Turns Agent Traces and User Actions into Skills, Evals, and Fine-Tuning Data Introduction Teams shipping AI agents at scale keep running into the same pattern. Even agents built on top-tier models like Fable and Astra…

  19. Mastodon — mastodon.social TIER_1 English(EN) · [email protected] ·

    CopilotKit 发布 CLiF,一个将代理跟踪和用户操作转化为技能、评估和微调数据的持续学习平台。该平台

    CopilotKit has released CLiF, a continual learning platform that turns agent traces and user actions into skills, evaluations and fine-tuning data. The platform aims to help AI agents improve continuously by learning from real user interactions. It is now available in early acces…