范式革命:人工智能基础理论源头创新的必由之路  

Paradigm Revolution:The Only Way for the Original Innovation of the Basic Theory of Artificial Intelligence

作　　者：钟义信[1] Zhong Yixin

机构地区：[1]北京邮电大学人工智能学院

出　　处：《学术前沿》2021年第23期20-40,共21页Frontiers

基　　金：国家自然科学基金项目“机制主义方法与高等智能理论”,项目编号为60873001;国家社科基金重大项目“信息哲学的历史、现状与未来”的研究成果,项目编号为18ZDA027。

摘　　要：人工智能是开放复杂的信息系统,应当遵循信息学科的范式。然而,由于受到"社会意识滞后于社会存在"法则的制约,信息学科范式至今未能形成,于是人工智能研究实际遵循的是物质学科范式,致使人工智能研究一直受困于初级阶段,只能取得局部的浅层进展。唯有通过人工智能的范式革命建立信息学科范式,才能获得人工智能研究的源头创新和整体创新,进入人工智能发展的高级阶段。正是沿着这个思路,需要研究和总结信息学科范式,以此突破物质学科范式对人工智能研究的束缚,并在信息学科范式引领下,发现全新的人工智能研究模型,揭示"信息转换与智能创生定律"这个普适的智能生成机制,创建源头创新和整体创新的人工智能理论——机制主义通用人工智能理论。Artificial intelligence (AI) is an open and complex information system,which should follow the paradigm of information science.However,due to restriction of the law that "social consciousness lags behind social existence",the information discipline paradigm has not been formed so far,so the AI research actually follows the material discipline paradigm,resulting in AI research being stuck in the primary stage and only making shallow progress in certain areas.Only by establishing the information discipline paradigm through the paradigm revolution of AI can we obtain the original innovation and overall innovation of AI research and enter the advanced stage of AI development.According to this idea,it is necessary to study and summarize the para-digm of information discipline,so as to break through the shackles of the paradigm of material discipline on the AI research;and under the guidance of the paradigm of information discipline,find a new AI research model,reveal the universal intelligent generation mechanism of the "law of information transformation and intelligent creation",and create an AI theory of original in-novation and overall innovation–the mechanism-based general AI theory.

关 键 词：人工智能 范式革命 机制主义 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]