（二十八）

报告题目: Complex Network Analysis of Structures in Mammalian Sensory Neural Networks

报 告 人: Edgar Fuller（美国西弗吉尼亚大学, 教授）

报告时间: 2018年5月9日(周三)上午9:00-10:00

报告地点: 安徽大学磬苑校区数学科学学院H306
（二十九）

报告题目: Nowhere-zero 11-flow of signed graphs

报 告 人: 罗荣（美国西弗吉尼亚大学, 教授）

报告时间: 2018年5月9日(周三)上午10:00-11:00

报告地点: 磬苑校区数学科学学院H306

主办单位：安徽大学数学科学学院

欢迎各位老师、同学届时前往！

科学技术处

2018年4月23日
报告摘要（二十八）:

Complex networks arise in contexts ranging from social network interactions to financial transactions to protein interaction structures within the human body. Once a network structure is identified, the tools of graph theory and network analysis can be used to identify important portions of the network including essential nodes and subnetworks of affiliated nodes. Within the field of neuroscience, neurons form networks within animal brains, motor control and sensory systems. Studies of the connectivity within the neural structures found in these areas can shed light on the ways in which information is processed and stored in these structures. This processing enables response to sensory input, memory and motor control among other functions.Work focusing on the identification of important structures within connectomes has primarily identified nodes in these networks that possess graph theoretic properties such as high connectivity or high betweenness (Towlson, et al 2013; Van Den Heuvel and Sporns 2011) and used combinations of network properties along with other methods to study cell types within these networks (Bullmore and Bassett 2011, Pailthorpe 2016). In this talk I will present a review of some of this work and then describe our effort to use data from Helmstaedter et al, (2013) and an improved version of the Quasi-Clique Merger community detection algorithm (Qi et al. 2012, 2014, 2015; Wu et al. 2013, Zhao and Zhang 2011) to identify structural communities of neurons within the inner plexiform layer (IPL) of the mouse retina. This process autonomously finds multiple levels of associations within neuron connectivity developed from SBEM imagery that agree with the biological structures found in previous work (Helmstaedter, et al 2013). We identify retinal computation circuitry subgroups (Briggmann, Helmstaedter and Denk 2011; Borst and Helmstaedter 2015; Greene, Kim and Seung 2016; Behrens, et al 2016) as community structures that are both densely connected within these subgroups as well as sparsely connected across groups to higher level structures such as ganglion cells and then use those relationships to deduce neuron function.

报告摘要（二十九）:

In 1983, Bouchet proposed a conjecture that every flow-admissible signed graph has a nowhere-zero 6-flow. The best published result is Zyka's 30-ow and was recently improved to 12-flow by DeVos. It Chen et al. further strengthened DeVos' result for bridgeless signed graphs, and show that every bridgeless flow-admissible signed graph admits a nowhere-zero 11-flow. Recently we show that the 11-flow theorem is true for all flow-admissible signed graphs.
报告人简介:

Edgar Fuller,美国西弗吉尼亚大学（West Virginia University，USA）数学系系主任，教授，IEEE会员。University of Georgia博士，主要研究方向包括微分几何学，数学教育，数据挖掘和生物信息。长期主持NASA和美国NSF项目。

罗荣，美国西弗吉尼亚大学（West Virginia University，USA）数学系教授。主要研究图的染色理论和流的理论，是国际知名的染色问题专家。发表50余篇论文，多数是发表在图论顶尖杂志上，如Journal of Cominatorial Theory Ser. B, Journal of Graph Theory, SIAM Journal on Discrete Math, European J. of Combinatorics. 在上世纪60年代末Vizing提出的四个关于边染色猜想的研究方面取得了一系列突破性进展。解决了几个著名的公开问题，如Erd?s、Gould Jacobson以及 Lehel 提出的一个关于可图序列猜想，Borodin 提出的边面染色的问题，以及Archdeacon 关于三流可图序列的问题。