AI论文速递 2026年08月13日（HuggingFace Daily Papers）

数据来源：https://huggingface.co/papers 采集时间：2026-08-13

📌 重点关注

1. SPIEval: Evaluating Large Language Models as Mobile Assistants over Scattered Personal Information | arXiv — 【重点关注】 Large language models (LLMs) are increasingly deployed as mobile assistants, ... 💡 移动端Agent评测基准，正对你的安卓/鸿蒙方向
2. ComBodied Agents: a New Paradigm of Human-Centric Agentic AI | arXiv — 【重点关注】 After an older adult misses a medication dose, a software agent can send anot... 💡 以人为中心的Agent新范式，健康场景可借鉴
3. MBA: Multimodal Benchmark and Agents for Real-World Business Ideation | arXiv — 【重点关注】 Agentic systems powered by large language models (LLMs) have opened new oppor... 💡 多模态商业构思Agent，启发AI提效工具

📋 其他值得关注

1. On-Policy Self-Distillation without Any Supervision | arXiv — On-policy (Self-)Distillation (OPD / OPSD) has shown strong potential for pos...
2. The Next Screenshot Knows: Gated Hindsight Distillation for Mobile GUI Agents | arXiv — GUI agents are commonly trained offline from successful interaction trajector...
3. VectraYX-Vision-1B: A Sub-2B Spanish/LATAM Cybersecurity Vision-Language Model with Structured Visual Reasoning and Native Tool Use | arXiv — We present VectraYX-Vision-1B, a sub-2B vision-language model (VLM) for Spani...
4. DistilVDR: A Compact End-to-End Visual Document Retriever via Dual-Student Distillation | arXiv — Visual document retrieval (VDR) is dominated by multi-billion-parameter model...
5. Spark-to-Paper: End-to-End Research Paper Generation as a Composable Skill | arXiv — Turning a research idea into a complete paper requires more than text generat...
6. ToolHazard: Scaling Adversarial Environments for Security Evaluation and Alignment of LLM-based Agents | arXiv — Large language model (LLM) agents integrated with external tools are vulnerab...
7. SkillZip: Evaluation-Free Skill Compression for Self-Evolving Agents by Discovering Reusable Structure | arXiv — Self-evolving agents accumulate reusable skills by appending successful proce...