arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2511.09149cs.LGcs.AIcs.MA
收藏

使智能体能够在潜在空间中完全交流

Enabling Agents to Communicate Entirely in Latent Space

State Key Lab of CAD&CG(CAD与CG国家重点实验室);Future Living Lab of Alibaba(阿里巴巴未来生活实验室);Zhejiang Key Laboratory of Medical Imaging Artificial Intelligence(浙江医学影像人工智能重点实验室);Shanghai Jiao Tong University(上海交通大学)

Zhuoyun Du, Runze Wang, Huiyu Bai, Zouying Cao, Xiaoyong Zhu, Yu Cheng, Bo Zheng, Wei Chen, Haochao Ying

AI总结:

本文提出Interlat方法,通过利用LLM的连续隐藏状态实现智能体间的潜在空间通信,实验表明其在跨异构模型中表现优异,提升了推理速度并保持性能。

摘要:

尽管自然语言是基于LLM的智能体的事实性通信媒介,但它存在根本性限制。将丰富的内部潜在状态下采样为离散token的过程本质上限制了信息传输的深度和细微差别,从而阻碍了协作问题解决。受telepathy启发,我们提出Interlat(跨智能体潜在空间通信)范式,利用LLM的连续最后隐藏状态作为其思维的表示,实现直接通信(称为潜在通信)。额外的学得压缩过程通过潜在空间推理进一步压缩潜在通信。实验表明,Interlat在跨异构模型中优于微调的链式思维(CoT)提示和单智能体基线,促进更具探索性的行为,并使潜在信息得到真正利用。进一步压缩不仅使推理速度提升高达24倍,还通过高效的保信息机制保持竞争性性能。我们将这项工作定位为完全潜在空间跨智能体通信的可行性研究,我们的结果突显了其潜力,为未来研究提供了有价值的见解。我们的代码可在https://github.com/XiaoDu-flying/Interlat获取。

英文摘要:

While natural language is the de facto communication medium for LLM-based agents, it presents a fundamental constraint. The process of downsampling rich, internal latent states into discrete tokens inherently limits the depth and nuance of information that can be transmitted, thereby hindering collaborative problem-solving. Inspired by telepathy, which bypasses symbolic language in communication, we propose Interlat (Inter-agent Latent Space Communication), a paradigm that leverages the continuous last hidden states of an LLM as a representation of its thought for direct communication (termed latent communication). An additional learned compression process further compresses latent communication via latent space reasoning. Experiments demonstrate that Interlat outperforms both fine-tuned chain-of-thought (CoT) prompting and single-agent baselines, even across heterogeneous models, promoting more exploratory behavior and enabling genuine utilization of latent information. Further compression not only substantially accelerates inference by up to 24 times but also maintains competitive performance through an efficient information-preserving mechanism. We position this work as a feasibility study of entirely latent space inter-agent communication, and our results highlight its potential, offering valuable insights for future research. Our code is available at https://github.com/XiaoDu-flying/Interlat.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑