arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2202.03153cs.AI
收藏

通用人工智能研究方法分析

Approaches to Artificial General Intelligence: An Analysis

Soumil Rathi

摘要:

本文分析了实现AGI的不同方法,包括人脑仿真、AIXI和集成认知架构。首先,本文对AGI的定义进行了界定,并阐述了AGI的要求。对于所提到的每一种拟议方法,都总结了所讨论的方法,并详细介绍了其关键过程,展示了其如何运作。然后,对每种方法进行了分析,考虑到各种因素,如技术要求,计算能力,以及是否符合要求。得出的结论是,虽然有各种方法可以实现AGI,如人脑仿真和集成认知架构,但实现AGI最有前途的方法是集成认知架构。这是因为人脑仿真被发现需要扫描技术,而这些技术很可能要到21世纪30年代才能获得,因此在此之前不太可能创建。此外,集成认知架构降低了计算需求,并为通用智能提供了合适的功能,使其成为实现AGI的最有可能的方式。

英文摘要:

This paper is an analysis of the different methods proposed to achieve AGI, including Human Brain Emulation, AIXI and Integrated Cognitive Architecture. First, the definition of AGI as used in this paper has been defined, and its requirements have been stated. For each proposed method mentioned, the method in question was summarized and its key processes were detailed, showcasing how it functioned. Then, each method listed was analyzed, taking various factors into consideration, such as technological requirements, computational ability, and adequacy to the requirements. It was concluded that while there are various methods to achieve AGI that could work, such as Human Brain Emulation and Integrated Cognitive Architectures, the most promising method to achieve AGI is Integrated Cognitive Architectures. This is because Human Brain Emulation was found to require scanning technologies that will most likely not be available until the 2030s, making it unlikely to be created before then. Moreover, Integrated Cognitive Architectures has reduced computational requirements and a suitable functionality for General Intelligence, making it the most likely way to achieve AGI.

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。

↑