AI News · 2026-10-01

按采集日期(UTC)汇总已发布资讯,原文日期单独标注。

返回资讯列表

Experience-Driven Continual Learning of Terrain Traversability for Quadruped Robots

arXiv cs.RO ·

Safe and efficient quadruped navigation over unfamiliar terrain requires predicting terrain-robot interaction before contact: geometry and visual appearance alone cannot reveal …

原文内容 · 暂无该语言译文

阅读原文 ↗

VehicleArena: A Realistic Urban Environment for Multi-Agent Driving

arXiv cs.CL ·

Real-world embodied agents often pursue independent objectives within a shared physical environment, where their actions can alter the conditions faced by others. Existing bench…

原文内容 · 暂无该语言译文

阅读原文 ↗

OpenRouter 教程:如何在 CI 中用 LLM eval 门禁拦截 Pull Request

AIHOT ·

OpenRouter 发布教程,讲解如何用固定的 eval 集在 CI 中门禁 pull request,当通过率低于阈值时脚本以非零退出码阻止合并,做法与单元测试门禁一致。

原文内容 · 暂无该语言译文

阅读原文 ↗

OpenAI 披露并处置一起有组织的模型蒸馏攻击行动

AIHOT ·

OpenAI 披露其识别并处置了一起有组织的攻击行动,该行动旨在系统性提取模型受保护的推理内容,最早活动出现在 7 月第一周。

原文内容 · 暂无该语言译文

阅读原文 ↗

One Model Is Not a Crowd: Multi-LLM and Aspect-Conditioned Diverse Comment Generation

arXiv cs.CL ·

Human communication on the internet is shaped by diverse perspectives, most visibly expressed in online comment spaces. As large language model (LLM)based AI agents begin to inh…

原文内容 · 暂无该语言译文

阅读原文 ↗

RynnWorld-4D: 4D Embodied World Models for Robotic Manipulation

arXiv cs.RO ·

Robotic manipulation in the open world requires not only recognizing what a scene looks like, but also anticipating how its 3D structure moves under interaction. We argue that s…

原文内容 · 暂无该语言译文

阅读原文 ↗