摘要
机制主义人工智能理论是基于智能的生长机制而把结构主义、功能主义和行为主义这三大人工智能流派有机统一起来并使基础意识、情感、理智成为三位一体的高等人工智能理论。因素空间是机制主义人工智能理论的数学基础,是现有模糊集、粗糙集和形式背景理论的进一步提升,它为信息描述提供了一个普适性的坐标框架,把数据变成可视的样本点,形成母体背景分布,压缩为背景基,由此进行概念自动生成,因果关联分析,以及建立在其上的学习、预测、识别、控制、评价和决策等一系列数学操作活动。本文将着重介绍其中的核心内容,将具体的形式信息(即语法信息)与效用信息(即语用信息)关联起来,提升为抽象的语义信息,为机制主义人工智能的信息转化第一定律提供一个简明的数学架构。本文以"九宫棋"为例,介绍如何用因素思维实现目标因素与场景因素的对接和搜索,为信息转化的第二定律从数学上展开探索性的思考;还结合因素空间及有关学科的历史来进行解说,以便帮助读者对因素空间理论有一个较为全面的认识。
Based on using the intelligent growth mechanism, the mechanism-based artificial intelligence theory organically unifies the structure, function, and behaviorism of three genres to form a trinity of consciousness, emotion, and reason. Factor space is the mathematical basis of mechanism-based artificial intelligence theory, which promotes mathematical branches such as formal concept analysis, rough sets, and fuzzy sets, and provides a universal coordinate framework for the description and cognition of things. Data can be represented as visual sampling points in the space and then be cultivated to form the population distribution of the background relation. Based on their relationship, concept generation and causality analysis can be performed automatically, and all rational thinking processes, such as prediction, identification, control, evaluation and decision making, can be performed by factorial algorithms. In this article, we focus on ways to describe formal information（i.e., grammatical information）, predict utility information（i.e., pragmatic information） from formal information, and correlate them to generate abstract semantic information, which is helpful for mathematically describing the first established law of information transformation in mechanism-based artificial intelligence theory. We also use factor space theory in chess Tic-Tac-Toe to demonstrate how to dock the target and chess factors,which may provide a clue for how to mathematically describe the second law of information transformation. We also provide a brief history to help readers gain a more comprehensive understanding of the factor space theory.
出处
《智能系统学报》
CSCD
北大核心
2018年第1期37-54,共18页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金委主任基金(61350003)
教育部高校博士学科点专项科研基金资助项目(20102121110002)
辽宁省教育厅科学技术研究一般基金资助项目(L2014133)
关键词
机制主义人工智能理论
因素空间理论
形式概念分析
粗糙集
模糊集
模糊落影理论
背景关系
数据挖掘
mechanism-based artificial intelligence theory
factor space theory
formal concept analysis
rough sets
fuzzy sets
falling shadow theory
background relation
datamining
