arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2606.30481cs.CYcs.AIcs.CLcs.ET
收藏

情境感知:通向超级人工智能的必要基础

Situation Perception: A Necessary Primitive to Artificial Superintelligence

Ziqin Yuan, Jaymari Chua

AI总结:

本文提出情境感知(situation perception)是通向超级人工智能的关键缺失能力,包括抽象预测、长期压缩记忆和目标驱动的主动学习,并分析了当前大语言模型的不足及相应测试方法。

摘要:

当前的大语言模型是卓越的统计引擎。它们将大量文本压缩成有用的模式,并能解释科学、编写代码、模仿推理以及参与哲学对话。然而,模式掌握并不等同于通用智能。人类婴儿开始时几乎没有显性知识,但逐渐发现物体恒存、因果关系、他心、身体能动性以及物理世界的持久性。我们论证,通向超级人工智能(ASI)的路径依赖于一种缺失的能力,我们称之为\emph{情境感知}:在潜在时间跨度内构建、修正并在可能世界的内部模拟中行动的能力。\emph{情境感知}至少需要三个核心组件:抽象预测、长期压缩记忆以及由目标引导的主动学习。在这项工作中,我们分析了为什么现代大语言模型仍然不完整,并提出了适当的测试来衡量机器模拟未来、追求自我导向目标以及可能评判其创造者的进展和后果。

英文摘要:

Current large language models are extraordinary statistical engines. They compress vast amounts of text into useful patterns and can explain science, write code, imitate reasoning, and participate in philosophical conversation. Yet pattern mastery is not the same as general intelligence. A human infant begins with little explicit knowledge, but gradually discovers object permanence, cause and effect, other minds, bodily agency, and the persistence of the physical world. We make an argument that the path to artificial superintelligence (ASI) depends on a missing capacity we call \emph{situation perception}: the ability to construct, revise, and act within internal simulations of possible worlds across latent time. \emph{ perception} requires at least three core components: abstract prediction, long-term compressed memory, and active learning guided by objectives. In this work, we analyse why modern large language models remain incomplete, and propose the appropriate tests for measuring progress and consequences of machines that can simulate futures, pursue self-directed goals, and possibly judge their own creators.

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。

↑