Skip to content

AI论文速递 2026年07月24日（HuggingFace Daily Papers）

数据来源：https://huggingface.co/papers 采集时间：2026-07-24

📌 重点关注

  1. NVIDIA-labs OO Agents: Native Python Object-Oriented Agents | arXiv — 【重点关注】 Traditional agent development is split across prompt templates, tool schemas,... 💡 用Python OOP替代prompt模板定义Agent，让Agent代码像普通程序一样可维护可测试，和你的Agent实践方向直接相关。
  2. An Exam for Active Observers | arXiv — 【重点关注】 Human vision is a closed loop: gaze is continuously redirected by intermediat... 💡 主动感知闭环研究，对端侧AI的视觉感知架构设计有启发。
  3. Computational Humor with Multimodal LLMs: Methods, Datasets, Evaluation, and Challenges | arXiv — 【重点关注】 Multimodal humor in memes, cartoons, and comics remains difficult for AI syst... 💡 多模态幽默理解是AI情商的关键测试，研究价值高但偏前沿。

📋 其他值得关注

  1. ConsiSpace: Learning Geometric Consistency Matters for Video Spatial Reasoning | arXiv — Video spatial reasoning is essential for navigation-oriented perception and l...
  2. SeededGrasp: Language-Guided Grasping in Complex Scenes with Multiple Embodiments | arXiv — Practical robotic grasping in complex scenes requires both 3D spatial reasoni...
  3. ICAE-Bench: Evaluating Coding Agents as Interactive Project Builders | arXiv — The recent emergence of vibe-coding workflows is changing what coding agents ...
  4. AutoIndex: Learning Representation Programs for Retrieval | arXiv — We present AutoIndex, a framework for learning representation programs: execu...
  5. AREX: Towards a Recursively Self-Improving Agent for Deep Research | arXiv — Deep research requires agents to find answers that jointly satisfy multiple c...
  6. SLAI T-Rex: Full-Parameter Post-training of the DeepSeek-V4 Family on Ascend SuperPOD | arXiv — Full-parameter post-training of trillion-parameter-scale MoE models introduce...
  7. Beyond Relevance-Centric Retrieval: Rubric-Oriented Document Set Selection and Ranking | arXiv — As large language models and AI agents become the primary consumers of search...