论文部分内容阅读
类比推理技术对于开发高智能的ＡＩ系统十分重要．普通的面向数值处理的计算机进行类比推理效率往往很低，严重影响其研究与应用．本文介绍了一种高效的嵌入式类比推理协处理器系统ＡＲＣ（ａｎａ－ｌｏｇｉｃａｌｒｅａｓｏｎｉｎｇｃｏｐｒｏｃｅｓｓｏｒ）．由于ＡＲＣ采用了符合类比推理机制的处理器体系结构和加速策略，并采用先进的器件实现，其推理速度比普通的ＰＣ机有大幅度的提高
Analogical reasoning is very important for the development of highly intelligent AI systems. Ordinary numeric-oriented computer analogy reasoning efficiency is often low, seriously affecting its research and application. This article describes an efficient embedded analogical reasoning coprocessor ARC (ana-logicalreasoningcoprocessor). Because ARC uses a processor architecture and acceleration strategy that conforms to analogical reasoning and is implemented using advanced devices, the deduced speed is significantly higher than that of an ordinary PC