管理员登录 / English
Fuzzy Neural Networks: At the Junction of Neurocomputing and Fuzzy Sets
发布时间: 2016-12-31     18:46   【返回上一页】 发布人:Witold Pedrycz


   北京师范大学数学科学学院

 

周五公众报告

 

 

 

报告题目:Fuzzy Neural Networks: At the Junction of Neurocomputing and Fuzzy Sets 

报告人:Witold Pedrycz 教授  (加拿大阿尔伯塔大学)

时间地点:201716日下午3:00-4:00, 后主楼1220

 

邀请人:于福生 教授

 

报告摘要: Neural networks and fuzzy sets constitute two fundamental technologies of Computational Intelligence. While both of them demonstrate some evident advantages, they also come with unavoidable limitations. Given the learning abilities of neural networks and strong knowledge representation aspects residing with fuzzy sets, their synergistic usage becomes instrumental in the alleviation of the limitations of the individual technologies.  

 

In this talk, we discuss a concept of fuzzy logic neurons and fuzzy logic networks, which are constructed as hybrid architectures. Individual neurons realize computing carried out in terms of logic operations (and and or connectives). Taxonomy of logic neurons embracing aggregation and referential computing units implies an emergence of a variety of neural architectures.  With this regard, a two-layer topology of the networks, being referred to as logic processors, is introduced and analyzed. Detailed learning schemes are discussed. Thorough design guidelines are developed with emphasis focused on various ways of accommodating prior domain knowledge and its calibration through parametric learning. A crucial interpretability facet of fuzzy neural networks is elaborated on.

 

Some representative application studies involve two essential categories of constructs such as rule-based models and pattern classifiers (with all of them represented in the form of fuzzy neural networks). 

 

 

 

报告人简介: 加拿大阿尔伯塔大学电子与计算机工程系的教授,加拿大智能计算首席首席科学家(Canadian Research Chair),该领域国际著名科学家。2009年,当选波兰科学院外籍院士;2012年当选加拿大皇家科学院院士。他是IFSA Fellow, IEEE Fellow。研究方向涉及计算智能、模糊建模与粒度计算、数据挖掘、模糊控制、模式识别、知识神经网络等领域,发表了大量学术论文,研究专著15本。Information SciencesIEEE Transactions on Systems, Man,and Cybernetics - Part A等国际著名杂志主编,IEEE Transactions on Fuzzy SystemsIEEE Transactions on Neural Networks等多个国际著名杂志编委。2007年获IEEESMCS协会颁发的最高技术成就奖IEEE Norbert Wiener Award2009年获Soft Computing领域国际最高奖Cajastur Prize,并获IEEE加拿大计算机工程勋章等。曾任国际模糊系统联合会(IFSA)和北美模糊系统协会(NAFIPS)主席。

 

周五学院公众报告主旨:讲解现代数学中的基本概念、重要结果及其数学思想的起源、方法的创新等,扩展我们的年青教师和研究生的视野、提高数学修养以及增强学院的学术氛围。邀请各方向的专家用通俗易懂的报告内容、自由互动的讲解方式展现现代数学中重要的基本概念和深刻原理,让听众可以领略到现代数学理论其实是来源于一些大学或研究生阶段就熟知的古典数学中的想法和结论。希望这些报告有助于高年级的本科生、研究生和年青教师更全面的认识现代数学、享受数学之美。欢迎大家踊跃参与!