HuggingFace Daily Papers（社区热门论文）· HuggingFace Daily Papers（社区热门论文）·· 8 天前AI 评分42
LLM 是通用异步智能体：Qwen 3.x 无需任务专项训练即可异步运行
AI 导读
研究提出异步 LLM 框架，让用户或智能体自身定义带重叠内存状态的推理协程，从而把 LLM 从"读取—思考—回复"的顺序交互推广到通用异步智能体。实验显示 Qwen 3.x 模型无需任务专项训练，即可完成流式视频理解、电子游戏和监控等异步任务。
正文
Abstract:Modern LLMs are increasingly capable as autonomous agents, but they follow sequential interaction cycles: read, think, reply or call tools, repeat. Many real-world use cases are not sequential: voice assistants, embodied agents, and monitoring systems receive new inputs while they think or perform another task. Modern LLMs address this with specialized architectures for voice interaction and video streams, VLAs for robot control, asynchronous tool calling for API usage, and others. In this work, we generalize from different asynchronous tasks to general asynchronous agents that can adapt to different types of concurrency. To achieve this, we develop an asynchronous LLM framework that lets users (or the agents themselves) define inference coroutines with overlapping memory states. We showcase that Qwen 3.x models are capable of asynchronous operation for streaming video understanding, videogames, and monitoring, without task-specific training.
| Comments: | Preprint |
| Subjects: | Machine Learning (cs.LG); Computation and Language (cs.CL) |
| Cite as: | arXiv:2609.35427 [cs.LG] |
| (or arXiv:2609.35427v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2609.35427 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: George Yakushev [view email]
[v1]
Mon, 28 Sep 2026 15:35:26 UTC (2,417 KB)
来源：HuggingFace Daily Papers（社区热门论文） · arxiv.org