arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2605.09271cs.AI
收藏

通过语言表示塑造模式作为扩展LLM智能的下一个前沿

Shaping Schema via Language Representation as the Next Frontier for LLM Intelligence Expanding

The Hong Kong University of Science and Technology(香港理工大学);MiLM Plus, Xiaomi Inc(小米公司);Zhongguancun Academy(中关村学院);Hong Kong Baptist University(香港 Baptist大学);The University of Tokyo(东京大学);RIKEN Center for Advanced Intelligence Project(日本理化学研究院高级智能项目中心);Xi’an Jiaotong University(西安交通大学)

Zhiqin Yang, Yuhan Liu, Jingwen Fu, Pei Fu, Bo Han, Masashi Sugiyama, Nanning Zheng

AI总结:

本文探讨通过高级语言表示塑造模式作为扩展LLM智能的新方向,提出语言表示的结构和符号复杂性对知识激活和组织至关重要,并通过实验证明语言表示设计的有效性。

摘要:

尽管自然语言是大型语言模型(LLM)的默认媒介,但其有限的表达能力为复杂问题解决带来了深刻瓶颈。尽管最近的AI进展主要依赖于扩大规模,但仅仅内部化知识并不能保证其有效应用。本文将语言表示定义为用于映射和建模现实世界的语言和符号构造,并论证通过高级语言表示塑造模式是扩展LLM智能的下一个前沿。我们提出,LLM的知识激活和组织——其模式——在很大程度上依赖于用于表示给定任务的语言的结构和符号复杂性。本文既提出了这一主张的正式化,也提供了支持其的实证证据。通过新的正式化,我们呈现了多个证据线支持我们的观点:首先,我们回顾了最近的经验实践和新兴方法,展示了通过有意识的语言表示设计可以实现显著的性能提升,即使不修改模型参数或规模。其次,我们进行了受控实验,显示在相同的基础任务下,LLM的性能及其内部特征激活在不同的语言表示下会有所不同。这些发现突显了语言表示设计作为未来研究有前途的方向。

英文摘要:

Although natural language is the default medium for Large Language Models (LLMs), its limited expressive capacity creates a profound bottleneck for complex problem-solving. While recent advancements in AI have relied heavily on scaling, merely internalizing knowledge does not guarantee its effective application. Defining language representation as the linguistic and symbolic constructs used to map and model the real world, this paper argues that shaping schemas through advanced language representation is the next frontier for expanding LLM intelligence. We posit that an LLM's knowledge activation and organization -- its schema -- depends heavily on the structural and symbolic sophistication of the language used to represent a given task. This paper contributes both a formalization of this claim and the empirical evidence to support it. With a new formalization, we present multiple lines of evidence to support our position: Firstly, we review recent empirical practices and emerging methodologies that demonstrate the substantial performance gains achievable through deliberate language representation design, even without modifying model parameters or scale. Secondly, we conduct controlled experiments showing that LLM performance and its internal feature activations vary under different language representations of the same underlying task. Together, these findings highlight language representation design as a promising direction for future research.

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。

↑