AI论文速递 2026年04月30日（HuggingFace Daily Papers）¶
数据来源：https://huggingface.co/papers 采集时间：2026-04-30
📌 重点关注¶
- GoClick: Lightweight Element Grounding Model for Autonomous GUI Interaction | arXiv — 【重点关注】 Graphical User Interface (GUI) element grounding (precisely locating elements...\n💡 轻量级GUI交互模型，降低AI界面理解门槛
- BARRED: Synthetic Training of Custom Policy Guardrails via Asymmetric Debate | arXiv — 【重点关注】 Deploying guardrails for custom policies remains challenging, as generic safe...\n💡 通过辩论训练安全护栏，强化AI政策合规性
- Vision-Language-Action Safety: Threats, Challenges, Evaluations, and Mechanisms | arXiv — 【重点关注】 Vision-Language-Action (VLA) models are emerging as a unified substrate for e...\n💡 多模态安全框架，AI系统安全保障
📋 其他值得关注¶
- Discovering Agentic Safety Specifications from 1-Bit Danger Signals | arXiv — Can large language model agents discover hidden safety objectives through exp...
- Recursive Multi-Agent Systems | arXiv — Recursive or looped language models have recently emerged as a new scaling ax...
- For-Value: Efficient Forward-Only Data Valuation for finetuning LLMs and VLMs | arXiv — Data valuation is essential for enhancing the transparency and accountability...
- AutoResearchBench: Benchmarking AI Agents on Complex Scientific Literature Discovery | arXiv — Autonomous scientific research is significantly advanced thanks to the develo...
- Efficient Agent Evaluation via Diversity-Guided User Simulation | arXiv — Large language models (LLMs) are increasingly deployed as customer-facing age...
- Taming Actor-Observer Asymmetry in Agents via Dialectical Alignment | arXiv — Large Language Model agents have rapidly evolved from static text generators ...
- DV-World: Benchmarking Data Visualization Agents in Real-World Scenarios | arXiv — Real-world data visualization (DV) requires native environmental grounding, c...