Augmenting Von Neumann's Architecture for an Intelligent Future
Rajpreet Singh, Vidhi Kothari
本研究提出了一种新型计算机架构，通过引入专用推理单元(Reasoning Unit, RU)扩展冯·诺伊曼模型，使系统具备原生通用人工智能能力。RU作为专用协处理器，将符号推理、多智能体协调和混合符号-神经计算作为基础架构原语执行。这种硬件嵌入式方法使得自主智能体能够在系统层面直接执行目标导向规划、动态知识操作和自省推理。该架构包含专为推理设计的指令集架构、并行符号处理流水线、智能体感知的内核抽象，以及无缝整合认知与数值工作负载的统一内存层次结构。通过跨越硬件、操作系统和智能体运行时层的系统化协同设计，该架构建立了推理、学习和适应作为内在执行特性（而非软件抽象）的计算基础，有望推动通用智能机器的发展。
This work presents a novel computer architecture that extends the Von Neumann model with a dedicated Reasoning Unit (RU) to enable native artificial general intelligence capabilities. The RU functions as a specialized co-processor that executes symbolic inference, multi-agent coordination, and hybrid symbolic-neural computation as fundamental architectural primitives. This hardware-embedded approach allows autonomous agents to perform goal-directed planning, dynamic knowledge manipulation, and i...
