《人工内分泌情感模型研究(1)》由会员分享，可在线阅读，更多相关《人工内分泌情感模型研究(1)（61页珍藏版）》请在金锄头文库上搜索。
1、中国科学技术大学 硕士学位论文 人工内分泌情感模型研究 姓名：梁建伟 申请学位级别：硕士 专业：计算机应用技术 指导教师：王煦法 2011-04 摘 要 I 摘摘 要要 近年来，受到生物信息处理机制的启发，人们提出了很多重要的智能模型和 技术。生物信息处理系统主要包括神经系统、免疫系统和内分泌系统，它们之间 不仅存在大的回路，而且具有直接的相互作用，对生物体的生长发育和新陈代谢 起着重要作用，共同维持着机体的正常生命活动。人们受到神经系统和免疫系统 生物机制的启发，已经分别提出了人工神经网络以及人工免疫系统，其理论、模 型、 算法和应用的成功研究充分显示出基于生物信息处理机制的人工智能模型研
2、究具有十分重要的理论意义和应用价值， 并促使人们对包括内分泌系统在内的其 他生物信息处理系统进行深入研究。 内分泌系统作为生物信息处理系统的重要组成部分， 蕴含着高度复杂和极其 独特的信息处理机制，主要包括激素调节机制、信息传递机制和情感反应机制。 抽取生物内分泌系统所独有的信息处理机制， 建立相应的模型算法具有重要的研 究价值，已经成为人工智能研究领域的热点问题。同时，情感作为人类智能中不 可分割的组成部分，在人的感知、推理、决策、创造和社交等诸多活动中起着不 可缺少的作用。随着脑科学、心理学、信息科学、智能科学的蓬勃发展，很多科 研工作者投入到人工情感、情感计算、虚拟现实等多个学科交叉研究
3、的领域，相 关理论及模型算法的研究对增强系统智能具有重要意义。研究发现，内分泌系统 和情感并不是孤立存在的，它们之间具有相互作用，内分泌系统对情感具有重要 的调节作用，而情感也会对内分泌系统造成难以忽略的影响。本文在深入研究和 分析生物内分泌系统的基础上，进行了人工内分泌情感模型研究。具体而言，本 文的主要工作包括： 一、从信息处理角度出发，研究生物内分泌系统的基本功能和作用机制，分 析和总结其独特的功能特点，为人工内分泌系统模型的研究提供理论依据和指 导。 二、在分析生物内分泌系统调节机制的基础上，提出一种基于激素调节的情 感模型。该模型中，内分泌系统对情感进行调节，能够避免智能体的情感状态
4、发 生突变，从而能够做出有效的行为决策，提高其在复杂动态变化环境中的生存能 力。仿真实验验证了模型及算法的有效性。 三、借鉴内分泌系统的情感反应机制，结合激素调节机制，引入和情感密切 相关的多种因素，从底层对生物体的心理和生理活动进行建模，本文首次提了出 一种内分泌系统和情感相互作用的人工内分泌情感模型。 考虑到激素分泌的复杂 性和情感本质上的模糊性， 采用了模糊逻辑的表征方法处理激素的分泌和情感的 产生过程。 摘 要 II 总的来说，生物内分泌系统蕴含着极其丰富的信息处理机制，在分析这些机 制的基础上，建立相关理论和模型具有重要意义。本文对人工内分泌模型进行了 初步研究，提出了人工内分泌情感
5、模型。 关键字：关键字：人工智能 人工内分泌系统 激素 情感 ABSTRACT III ABSTRACT For the past few years, inspired by the biological information processing mechanism, a lot of important intelligent models and technologies have been presented. The biological information processing system is mainly composed of nervous system, imm
6、une system and endocrine system. There is not only large loop among them, but also direct reciprocity, which plays an important role in the growing development and metabolism of organisms and maintains their vital movement together. Inspired by the biological mechanism of the nervous system and immu
7、ne system, artificial neural network as well as artificial immune system has been promoted. Successful research of the theories, models, algorithms and applications have shown the significant theoretic meaning and applied value of artificial intelligence models research based on biological informati
8、on processing mechanism, meanwhile ,prompted researchers to explore other information processing system of organisms including endocrine system. As an important part of biological information processing system, the endocrine system contains rather complex and particular information processing mechan
9、isms, including hormonal regulation mechanism, information delivery mechanism and emotion reacting mechanism. Extracting these mechanisms which are peculiar to endocrine system and establishing corresponding models and algorithms has important research value, which has been a hot issue in the resear
10、ch field of artificial intelligence. At the same time, as an integral part of human intelligence, emotion plays an essential role in perception, reasoning, decision making，creativity, social behavior, and so on. With the rapid development of brain science, psychology, information science and intelli
11、gent science, many researchers devote themselves into the research of artificial emotion, affective computing, virtual reality and other interdisciplinary fields. Theories, models and algorithms related to emotion are of great importance to enhance system intelligence. It has been discovered that en
12、docrine system and emotion dont exist in isolation, and there is reciprocity between them. Endocrine system plays a momentous regulatory role on emotion, and the great influence from emotion cannot be neglected. Based on the in-depth research and analysis of biological endocrine system, this paper d
13、id some research work on the artificial endocrine-emotion models. The main work is summarized as follows: ABSTRACT IV (1) From the information processing point of view, this dissertation lucubrates into the basic functions and mechanism of the endocrine system, analyzes and summaries its unique feat
14、ures, which offers theoretical basis and guidance for the research of artificial endocrine system model. (2) At a basis on the analysis of regulation mechanism of the biological endocrine system, the article brings forward an emotion model which is based on hormonal regulation. The endocrine system
15、regulates the emotions, which can avoid the occurrence of mutations of agents emotional state, leads to effective behavioral decision, and thereby improves the survivability in complex and dynamic changing environment. Experimental results have verified the validity of the supposed model and algorit
16、hm. (3) Using the emotion reacting mechanism of biological endocrine system, taking hormonal regulatory mechanism into consideration, introducing many factors which are closely linked to emotion, this dissertation brings forward an artificial endocrine-emotion model for the first time, in which the endocrine system and emotion have interaction. Fuzzy logic method is employed to deal with the complexity in hormone secretion and uncertainty in emotion generation, considering that the hormone sec
