学术讲座：Investigating Dynamic Networks

应国际学院的邀请，英国伦敦玛丽女王大学（QueenMary, University of London）理查德·克莱格（Richard Clegg）博士将为广大师生作题为《Investigating dynamic networks》的学术报告。

讲座题目:Investigatingdynamic networks

主讲人: Dr.Richard Clegg ——英国伦敦玛丽女王大学

时间: 2017年12月21日（周四）晚上18:00 -- 19:00

地点:西土城校区教三楼3-435

讲座内容:动态网络主要研究互连的网络节点随时间变化的特点和规律。凡是互相连接的对象集合均可视为网络：城市道路（通过街道连接的交通网络）、社会网络（例如Wechat，通过在线交互形成的个人网络）、因特网（通过电子链路构建的计算机网络）。

本讲座首先介绍各种网络之间的结构相似性（如朋友圈与大规模因特网在数学描述／含义上，具有类似的结构），然后讨论偏好连接（preferential attachment）法则，及基于该假设的网络增长／变化模型。偏好连接法则只是真实世界网络的一种近似，因此Richard Clegg博士将重点阐述其研究成果FETA（Framework for Evolving Topology Analysis,演进拓扑分析框架）在动态网络描述和分析方面的性能。

该讲座为前沿讲座，欢迎全校师生踊跃参加。

国际学院

2017年12月19日

附：

Short Bio of Dr. RichardClegg:

Richard G. Clegg is a Lecturer in Networks at Queen MaryUniversity of London. His PhD in mathematics and statistics from theUniversity of York was gained in 2005. His research interests include investigationsof the dynamic behaviour of networks and measurement of network trafficstatistics.

Brief introduction ofthis talk:

In this short talk I will introduce thestudy of dynamic networks, how networks of connected items evolve as time goeson. The networks can be any collection of connected objects. A city's roads canbe seen as a network of junctions connected by streets; A group of people canbe seen as individuals connected by friendships; A social network like Wechatcan be seen as individuals connected by their online interactions; The Internetcan be seen as a group of computer networks connected by electronic links.

I will show how many very different typesof networks can be shown to have a similar structure. A network of friends hasa similar structure mathematically to the large scale Internet. A law known as preferentialattachment has been hypothesised to cover how such networks might grow.However, this law is only approximate for the study of real world networks. Iwill show how my work on FETA (Framework for Evolving Topology Analysis)provides a way to introduce a flexible group of models of networks and also howit provides a reliable method to show which model is best to explain a givennetwork.