August 2018 - Dm-meeting - mailman.cs.vt.edu

Congratulations to Sorour!
by B. Aditya Prakash 29 Aug '18

29 Aug '18

to finishing her thesis proposal! Onwards towards the PhD :-) cheers, - Aditya

1 0

Preliminary exam: Sorour Amiri, Torg 3160A, 1-3pm, Aug 29 (Tomorrow)
by B. Aditya Prakash 28 Aug '18

28 Aug '18

Hello everyone, Sorry for the late notice: if you are free, please join us for Sorour Amiri's proposal (prelim exam) TOMORROW (Aug 29) 1pm at Torg 3160A. You can also join through zoom @ https://virginiatech.zoom.us/j/728249196 Meeting ID: 728 249 196 Title+abstract are below. Best, - Aditya ================== ***Title*** Task-specific summarization of networks: Optimization and Learning *** Abstract *** Networks (also known as graphs) are everywhere. People-contact networks, social networks, email communication networks, internet networks (among others) are examples of graphs in our daily life. The increasing size of these networks and simplicity to collect data diminishes our understanding of patterns in data. It is essential and at the same time challenging to analyze and understand these networks. Consider a social network such as Facebook or Twitter. A graph can represent the friendship between users. If we take the profile information or the strength of the friendship of users into account, we can have a weighted attributed graph. Also, we can consider the information associated with time stamps and produce a temporal friendship network. Mining and summarizing these graphs is an important task which reveals key patterns and helps in many applications such as visualizing and highlighting different communities or influential users in the network. In addition, summarizing networks helps in accelerating existing graph algorithms. We can run expensive algorithms such as “finding influential users” or “generating high-quality graph layouts” on smaller summary graphs and map the results to the original graph later. Finally, summarizing networks provides a high-level abstraction of them and helps in making generalized methods for a set of similar graphs. It is notable that different summarizes are desired for different purposes. For example, to stop viral infections, one may want to find an effective policy to immunize people in a people- contact network. In this case, a high-quality network summary should highlight the best people to immunize. Others may want to detect communities in the same people-contact network. Hence, we need another summary to show the communities. This means that for each task, we must design a specific method to reveal the related patterns. Despite the importance of summarizing networks for particular tasks, there has not been much work in this area. In this thesis, we design task-specific summarization frameworks for univariate and multi-variate networks to reveal their important patterns and to accelerate existing algorithms and making sense of them. We start with deterministic approaches to summarize graphs for a particular task and finally propose general frameworks which learn how to summarize the network for a given task and generalize it to similar networks. 1. Deterministic approach: Given a large network and a task, we propose summarization algorithms to highlight the characteristics of the graph (i.e., structure, attribute, label, dynamics) with respect to the task. We develop effective and efficient algorithms for summarizing univariate and multivariate graphs given various tasks such as content-aware influence maximization and time segmentation. In addition, we study many real-world networks and their summary graphs such as people-contact, news- blogs, etc. and visualize them to make sense of their characteristics given the input task. 2. Learning approach: As our next step toward graph summarization, we propose a unified framework which learns the process of summarization for a given task. First, we design a generative algorithm to learn to summarize graphs for any Graph Optimization Problem (GOP). Next, we go further and add sparse human feedback to the learning process of summarization for the given optimization task. Our models and frameworks lead to meaningful discoveries. We also solve problems from various domains such as epidemiology, biology, marketing, social media, and cybersecurity. *** Committee **** B. Aditya Prakash (chair) T. M. Murali Chris North Chandan Reddy Hanghang Tong (ASU)

1 0

Tuesday, Aug. 14th, 2018, @ 3:00 pm, McB 133 C, Sorour Amiri
by Sorour Ekhtiari Amiri 05 Aug '18

05 Aug '18

Hi everyone. Please join us for our next DM-Meeting. ****DM-Meeting Fall 2018**** WHO: Sorour E. Amiri WHEN: Tuesday, Aug. 14th, 2018, @ 3:00 pm WHERE: McB 133 C WHAT: I will give a prep talk for my preliminary exam. Title: Task-specific summarization of networks: Optimization and Learning Best regards, Sorour

1 0