报告人简介：

巩馥洲简介
现任数学与系统科学研究院应用数学研究所所长,中国科学院随机复杂结构与数据科学重点实验室主任，中国数学会副理事长。1986年毕业于西北大学，1989年于西北大学获硕士学位；1995年于中科院应用数学所获博士学位，师从严加安院士，后在北京大学数学系做博士后。研究方向是无穷维随机分析。曾获香港求是"杰出青年学者"奖和陈省身数学奖，国家杰出青年基金。

冯克勤简介
曾担任中国科技大学数学系主任，中国科技大学副校长，中国科学院研究生院常务副院长，清华大学数学科学系主任，信息安全国家实验室主任等职。1964年毕业于中国科学技术大学数学系；1968年中国科技大学数学系，研究生毕业，师从华罗庚教授。主要研究领域是代数数论和代数编码。曾获国家自然科学三等奖和陈省身数学奖。

王世坤简介
任中科院数学与系统科学研究院应用数学研究所研究员。1967年毕业于南开大学，1981年获硕士学位，导师是华罗庚教授和龚升教授。研究领域是数学物理和多复分析。曾获中科院科技进步一等奖和二等奖各一项、国家自然科学二等奖和三等奖各一项、中科院自然科学一等奖和二等奖各一项。

另外，我所安鸿志研究员28日18:00-20:00还将在数学院一层大报告厅做报告《“红学”中的数学思考》，欢迎对红学有兴趣的师生参加:-)

大偏差理论的历史可追溯到1930年代斯堪地那维亚的保险业。 在1960-1970年代中Ventcel, Freidlin, Donsker和Varadhan的工作使得这一理论得以飞速发展， 并且在很大的范围内对此理论产生极大的兴趣。从那以后至今大偏差理论已经在许多方向得到发展和应用。在此报告中我们将回顾大偏差理论从始至今的发展过程， 并着重介绍这段历史中一些重要理论突破。

Varadhan 教授于1959年获得印度Presidency College 学士学位，于1963年获得著名的印度统计学院 博士学位。 从1963 年起至今在美国纽约大学柯朗数学研究所任教。 Varadhan 教授现任 Frank Jay Gould 科学讲席教授。
Varadhan教授在偏微分方程理论， 大偏差理论，扩散过程理论及统计物理等众多领域都作出过基础开创性的贡献。Varadhan 的热核小时间渐近公式是偏微分方程理论中最著名的结果之一。 Varadhan 引理是大偏差理论的关键结果。 Varadhan 创建了扩散过程的鞅问题理论。这一理论是扩散过程在应用数学和纯粹数学中的应用中的基本工具。 Varadhan 的流体动力极限理论是现代粒子统计物理的数学基础。
Varadhan教授堪称是当今世界最有影响的概率学家。

部分荣誉及奖项：
美国科学院院士 （1995）
美国数学会 Leroy P. Steele 基础数学贡献奖 （1996）
阿贝尔数学奖 （2007）
印度总统Padma Bhushan 奖 （2008）

We provide a sufficient condition for the continuity of real valued permanental processes. When applied to the subclass of permanental processes which consists of squares of Gaussian processes, we obtain the sufficient condition for continuity which is also known to be necessary.Using an isomorphism theorem of Eisenbaum and Kaspi which relates Markov local times and permanental processes we obtain a general sufficient condition for the joint continuity of the local times. We show that for certain Markov processes the associated permanental process is equal in distribution to the loop soup local time..

Let X_1, X_2, .... be a sequence of i.i.d. symmetric random variables. I will prove that: (1) if X_1 has a density, then the first n partial sums S_1,S_2,...,S_n of X_1,X_2,...,X_n stay positive with a probability (2n-1)!!/(2n)!!; (2) if X_1 has a moment generating function, then the first n partial sums of S_1, S_2,...,S_n stay positive with a probability of the order n^{-1/4} . This is a joint work with Amir Dembo.

Suppose a sequence of independent normal random variables with unit variance are observed sequentially and the means are subject to abrupt change at an unknown change point. First, several detecting procedures are introduced along with their optimality properties. Second, the estimations (biases and confidence intervals) of the change point and post-change mean are considered. Third, extensions and applications to quality control, financial data monitoring, and other areas are used for illustration.

国家基金委医学部主任，肿瘤学、分子生物学专家，第二军医大学东方肝胆外科医院主任医师、教授，博士生导师，中国工程院院士。教育部“长江奖励计划”特聘教授，总后勤部“科技金星”。

中科院水生生物研究所所长，植物生理学及藻类学家，中国科学院院士。北京大学生命科学学院长江学者、特聘教授、常务副院长。

中国科学院北京基因组研究所所长，生物系统学和分子遗传学专家，台湾中央研究院院士，芝加哥大学教授,原进化和生态系系主任。

How to map the nodes and links efficiently in a given virtual network (VN) to those in the substrate network (SN) so that the residual substrate network (RSN) can host as many VN requests as possible is a major challenge in virtual network embedding. Early research has developed heuristic algorithms with interactive or two-stage methods. These methods, however, could cause the RSN fragmented into several disconnected components that are insufficient to host a large number of given VN requests. Without loss of generality, we assume that SNs, as the Internet, are small world and scale free, meaning that the average number of hops between any two nodes is a small constant and that most nodes have small degrees while only a small number of nodes have large degrees. Taking advantage of these two properties, we devise in this paper a new two-stage VN embedding approach to improve the connectivity of the residual substrate network so that it has the capacity to hostmore VN requests. Our algorithm uses a greedy strategy that maps neighboring VN nodes to substrate nodes whose distances are bounded by a small constant. We also map the edge nodes of given VN, i.e., nodes with small degrees, to nodes with large degrees in the SN. We then map the links by solving shortest path problems. Our experimental evaluations show that our algorithms offer better mappings and results in significantly fewer rejections for VN requests.

This is joint work with Yan Luo (UMass Lowell) and Xuzhou Chen (Fitchburg State College)

Dr. Jie Wang is Professor and Chair of the Department of Computer Science at the University of Massachusetts, Lowell, USA. He is also Director of the Center for Network and Information Security in the university and co-Director of Cyber Forensics Lab in the department. He received his PhD in Computer Science from Boston University in 1991, Master of Engineering in Computer Science from Zhongshan University in
1985, and Bachelor of Sciences in Computational Mathematics from Zhongshan University in 1982. His research interests include computational complexity theory, combinatorial optimization algorithms, computational medicine, and network security. He has worked as a security consultant in financial industry. His recent research focus is on network dynamics, knowledge discovery, and wireless sensor networks. His research has been funded by the NSF since 1991. IBM, Intel, Google and the Natural Science Foundation of China have also funded his research. He has published over 120 research papers in some of the most prestigious journals and conference proceedings. He has authored and co-authored three books; edited and co-edited four books. He is active in professional service, including chairing conference program committees, serving as journal editors, and organizing conferences and workshops.

随机微分几何与随机变分计算（现被称为Malliavin Calculus）创始人。

另有系列课程：
Ergodicity in Infinite Dimension and Stochastic Analysis
时间：2010年4月27日、28日 下午15:00开始

报告人简介：
http://www.bschool.nus.edu.sg/depart/ds/sunjiehomepage/