今日速览
| 序号 | 标题 | 来源 | 日期 | 主题 | 推荐等级 |
|---|---|---|---|---|---|
| 1 | Efficient Test-Time Adaptation through Human-AI Interaction | arXiv | 2026-09-03 | RAG | 高 |
| 2 | The Natural Language Interaction Protocol and Standard for AI Agents | arXiv | 2026-09-03 | MCP | 高 |
重点论文与技术动态
1. Efficient Test-Time Adaptation through Human-AI Interaction
- 来源：arXiv
- 日期：2026-09-03
- 作者/机构：Zora Zhiruo Wang, Apurva Gandhi, Rulin Shao, Aspen Chen, Jonas Mueller, Zhiqi Liang
- 主题标签：
RAG,arXiv - 推荐等级：高
- 分类：cs.AI
一句话结论
TAHI通过人机交互实现高效测试时适应，显著提升个性化任务成功率并支持跨用户泛化。
核心内容
- 提出测试时适应通过人机交互（TAHI），利用跨会话交互数据整合到代理上下文和权重中，并通过演变评分模块结晶用户训练和评估标准。
- 在写作和视觉创作领域适应30个个体，共600个任务，仅需数十任务即提升任务成功率4.5-20.9%，且个性化代理在跨用户上泛化成功率提升8.8%。
- 演变评分模块作为可扩展标注工具，创建评估标准捕捉16.0-22.3%更多失败，优于语言模型或人类单独评估。
方法与数据
- 适应30个个体在写作和视觉创作领域，共600个任务。
价值判断
- 值得关注：利用交互数据闭合个体专家差距，减少对详细规格的依赖，适用于开放性任务场景。
- 可复用点：演变评分模块可作为通用标注工具，自动创建评估标准，提升评估准确性和效率。
- 局限/待核查：摘要未明确。
摘要
AI agents are trained on population-scale data to encode broad capabilities spanning those of many practitioners. Yet the artifacts they produce rarely meet the personal bar professionals need to stake their reputation on. On realistic, open-ended tasks where success criteria are heterogeneous and insufficiently documented, individual expertise lives precisely in the elevation and departure from the average. In practice, iterative human-agent interaction surfaces criteria that users cannot fully specify up front, yet apply repeatedly across tasks. We argue this cross-session interaction data is a rich, underused signal for closing the gap to individual expertise. In this work, we propose test-time adaptation through human-agent interaction (TAHI), which integrates these signals into agent context and weights, and crystallizes each user's training and evaluation criteria via an evolving rubric module. We adapt agents to 30 individuals in two high-utility domains, writing and visual creation, on a total of 600 tasks. Our agents improve solo task success by 4.5-20.9% within only tens of tasks. Meanwhile, our evolving rubric module serves as a scalable annotation tool, creating evaluation rubrics that catch 16.0-22.3% more failures than those from LMs or humans alone. While agents are adapted towards individuals, we show these personalized agents also produce improvements in success of up to 8.8% that generalize across users.2. The Natural Language Interaction Protocol and Standard for AI Agents
- 来源：arXiv
- 日期：2026-09-03
- 作者/机构：Luyi Xing, Rasit Onur Topaloglu, Ranjan Sinha, Abhay Ratnaparkhi, Samuel Ndichu, Christopher Nguyen
- 主题标签：
MCP,arXiv - 推荐等级：高
- 分类：cs.AI
一句话结论
NLIP作为Ecma International标准化的协议，解决了AI代理在异构开发框架中的互操作性问题，通过轻量级语义消息封装支持HTTP/HTTPS、WebSocket和AMQP等传输协议，促进社会和商业应用落地。
核心内容
- NLIP由跨公司和大学的研究者开发，经Ecma International标准化，定义了AI代理交互的应用层协议。
- 提供轻量级语义消息封装，兼容HTTP/HTTPS、WebSocket和AMQP等传输协议，并包含安全设计考虑。
- 允许NLIP感知的代理和网关在客户端、本地上下文存储、本体、工具、企业服务及异构协议间无缝适应。
方法与数据
- 摘要未明确具体方法细节。
- 摘要未明确数据来源。
价值判断
- 值得关注：NLIP作为标准化协议，显著提升AI代理互操作性，解决异构环境通信问题，推动行业应用加速。
- 可复用点：协议设计兼容多种传输协议，便于集成到现有AI系统，支持跨平台应用开发。
- 局限/待核查：摘要未明确局限。