Presented by: Dr. Stacy Patterson
 
Abstract: Distributed estimation algorithms generate an estimate of a global network state using only local interactions. In large networks with a vast wealth of data, it is important that these algorithms capture relevant information in a compact form, and that they are robust to network dynamics.

In the first part of this talk, I will focus on the effects of network dynamics on a well-known distributed estimation problem, the distributed average consensus problem. Distributed average consensus algorithms have a wide variety of applications, including sensor fusion, distributed optimization, and autonomous vehicle formation control. I will present our recent work on the stability and robustness of consensus algorithms in dynamic networks and show the analytical relationship between algorithm performance and the network size, topology, and dynamic characteristics. A notable result is that network topology imposes fundamental limitations on the scalability of distributed consensus algorithms that cannot be overcome without global information.

In the second part of the talk, I will address a more sophisticated approach to distributed estimation based on compressed sensing. Recent works have demonstrated that compressed sensing is applicable to a variety of problems in sensor networks, including urban environment monitoring and traffic estimation. I will show how we leveraged work in the database literature to develop a distributed compressed sensing algorithm that outperforms previous approaches by several orders of magnitude in both time and message complexity, making it an efficient solution for resource-challenged networks. I will then connect this algorithm with distributed average consensus by showing how we combine the two algorithms to obtain an efficient distributed compressed sensing algorithm suitable for dynamic networks.
 
Bio: Stacy Patterson is a Technion Postdoctoral Fellow and a Viterbi Postdoctoral Fellow in the Department of Electrical Engineering at Technion – Israel Institute of Technology. She received her M.S. and Ph.D. in Computer Science from the University of California, Santa Barbara in 2003 and 2009, respectively. From 2009 to 2011, she was a postdoctoral scholar in the Center for Control, Dynamical Systems, and Computation at the University of California, Santa Barbara. Her research interests are in distributed algorithms and applications for aggregation, estimation, and control in dynamic networks.

Presented by Isaac Griffith
 
Abstract: Technical debt is a metaphor describing the economic consequesnces associated with taking short cuts to satisfy the short term goals of a project (i.e., meeting a release deadline) at the expense of long term goals (i.e., refactored software that meets quality criteria). Technical debt management is essentially the identification, measurement, and decision analysis surrounding the remediation and acceptance of technical debt in a project. In this presentation I describe previous research conducted in the areas of automated refactoring and code smell detection in the context of managing technical debt.  Further, I describe current research specifically dealing with the management decisions involving release planning in the context of technical debt managment. In conclusion, I will show how this previous research has chartered an initial research path towards my dissertation in design pattern evolution and design pattern grime.

Presented by: Ethan Katz-Bassett
 
Abstract: The Internet is now central to many aspects of modern society, yet it remains remarkably fragile. Partial outages are common, and performance problems are widespread. Operators would like to address these issues, but poor diagnostic tools hamstring their efforts. I will argue that a more robust Internet -- one with the predictable performance and high availability needed to provide critical services -- requires the development of a new generation of better tools. We must move towards a self-healing Internet that fixes problems rapidly, without requiring the hours or days that operators often currently take. With collaborators, I have developed practical distributed systems to understand Internet problems and to provide crucial steps towards automated remediation. Our systems are deployable today, without requiring modifications to the network. In the first half of the talk, I will present Reverse Traceroute, a system to measure the routing and performance behavior of reverse paths back to the local host from other networks. While tools have long existed to measure the forward direction, the reverse path has been largely opaque, hindering troubleshooting efforts. I will show how Google and other content providers can use reverse traceroute to troubleshoot their clients' performance problems. In the second half of the talk, I will present LIFEGUARD, our system that uses Reverse Traceroute and related techniques to diagnose and automatically repair availability problems, even without the participation of the network containing the failure.
 
Bio: Ethan Katz-Bassett is an Assistant Professor at the University of Southern California. Previously, he worked for Google investigating mobile performance and Internet routing. He earned his PhD in Computer Science at the University of Washington in 2012, where his dissertation won the department's William Chan Memorial Dissertation Award. His research also won Best Paper awards at NSDI in 2008 and 2010. Ethan's research goal is to design deployable techniques to dramatically improve Internet reliability and performance, based on the needs of operators and providers and grounded in rigorous network measurements. 

Presented by Dr. Rafal Angryk
In this talk Dr. Angryk will summarize recent research performed in the MSU’s Data Mining Laboratory. The focus of the entire seminar is on Big Data Analytics of Solar Dynamics Observatory (SDO), which is a flagship of NASA’s current “Living with a Star” program. The audience will first learn about the importance of solar data analysis, then about complexity of data maintained on the servers in our department. After that, three large research projects will be introduced with the hope of creating research interest among our graduate students. Finally, we will briefly talk about the future of the massive Solar Data Mining effort at MSU, and the potential for new research funding.