AI论文速递 2026年08月07日（HuggingFace Daily Papers）¶
数据来源：https://huggingface.co/papers 采集时间：2026-08-07
📌 重点关注¶
- SmartMage: Dynamic Modality Orchestration for 3D Scene Understanding | arXiv — 【重点关注】 Understanding 3D scenes is fundamental to embodied intelligence, requiring jo... 💡 动态多模态路由按需选择最有价值信息，对端侧AI多模态决策架构有参考意义
- K-EXAONE 2.0 Technical Report | arXiv — 【重点关注】 This technical report presents K-EXAONE 2.0, an open-weight multilingual foun... 💡 LG开源多语言大模型技术报告，东亚语言理解能力值得关注
- ToolArtist: Tool-Using Unified Multimodal Models for Agentic Image Generation | arXiv — 【重点关注】 Text-to-image (T2I) models can produce visually compelling images, yet they r... 💡 Agent+图像生成的统一策略，工具调用与原生生成一体化，生成式Agent新范式
📋 其他值得关注¶
- Learning from Failures: Retrieval-Centric CoT via Hard Negatives for Unified Multimodal Retrieval | arXiv — Unified multimodal retrieval aims to identify candidates that satisfy complex...
- ChronoVision: Temporal Reasoning via Latent State Reconstruction | arXiv — Multimodal large language models excel at passive perception but struggle wit...
- FocusMem: Factorizing Content, Readout, and Trust in Latent GUI Memory | arXiv — GUI agents must remember both useful experience from earlier tasks and unfini...
- BridgeVLA++: A Data-Efficient, Generalizable, and Memory-Augmented Vision-Language-Action Framework for 3D Manipulation | arXiv — Leveraging pre-trained vision-language models (VLMs) to construct vision-lang...
- OneDayAgent: Towards a Long-Horizon Harness for Autonomous Agents | arXiv — LLM agents are increasingly applied to open-ended everyday requests that span...
- EnvACE: Internalizing Environment Dynamics via World Rehearsal for Agentic Reinforcement Learning | arXiv — Training large language model agents for long-horizon tool use typically reli...
- HarnessOpt-Bench: Evaluating LLMs at Harness Optimization | arXiv — As LLMs are increasingly deployed within agentic systems, their capabilities ...