Interview with Dr. Yang Song

Featured Engineer

Dr. Yang Song

Dr. Yang Song - IBM T. J. Watson Research Center

What are your favorite hardware tools that you use?

Spectrum Analyzer + Signal Generator

What are your favorite software tools that you use?

MATLAB is my favorite tool to calculate and draw illustrative figures. Other than that, LaTeX is the software I use every day to write reports, papers, preparing presentation slides, resume, and keep track of my daily research thoughts. Better and prettier than scratch papers that will disappear mysteriously very soon.

What is on your bookshelf?

Well, most of books on my bookshelf right now are mathematic books which are very helpful in my networking research. For example, “Algebraic Graph Theory,” “Introduction to Graph Theory,” “Queueing Systems” help to model the communication network into mathematical abstractions; “Nonlinear Programming,” “Network Flows: Theory, Algorithms, and Applications,” “Combinatorial Optimization,” “Numerical Optimization,” “Introduction to Operation Research” are used to characterize the performance limits and tradeoffs of communication networks; “Introduction to Algorithms,” “Approximation Algorithms,” “Stochastic Approximation and Recursive Algorithms and Applications” enable me to design algorithmic solutions to networking problems, and to analyze how well they will perform.

Do you have any tricks up your sleeve? (special way to analyze circuits, special process you use to make something, etc.)

As a researcher, I constantly have no clue about whether the problem I am currently working on can be solvable or not. My way is to start from the simplest version of the problem and to explore from there. Say, if I want to address an optimization problem for a network with hundreds of nodes, I will first start from only two nodes, assuming all network information are known. If a solution can be found in this case, then a 3-node scenario will be considered, and so on. This would be a starting point from where I continue the journey by gradually relaxing the assumptions (often unrealistic) I made in simple cases. A thorough examination of simple cases can help me to understand the essence of the problem which will point out the right direction I should pursue further.

What has been your favorite project?

My favorite project is always the “current one”. My enthusiasm and curiosity on a challenging research problem grow exponentially with the time I have spent on it. After the work is done, however, this fascination vanishes linearly. Sometimes I feel that the problem itself appears less charming than it was at the initial stage, when I have absolutely no idea how to solve it.

What are you currently working on?

At IBM Research, I spend most of my time on wireless networking related research problems. Currently I am involved in a project jointly funded by UK and US to study network and information science. The goal of this project is to increase the interoperability of coalition networks with multiple group members. Due to different views of coalition groups, a consensus on how to utilize the network resource efficiently is difficult to achieve. I am currently developing algorithmic solutions that are able to optimize certain global network performance objectives without revealing the private and hence sensitive information of each coalition member. It is fun, really.

What challenges do you foresee in our industry?

As a wireless networking researcher, I feel that the driving force of wireless technology development is the increasing traffic demand from end users and new wireless applications. Data traffic becomes dominant over voice traffic. For example, it would be very difficult for me to imagine that I can watch online videos on my cellphone ten years ago. Now look around. People are even talking about playing online interactive video games on cellphones. The demands from new wireless applications impose significant challenges on current wireless technologies and infrastructures. Great. This means we have more fun problems to work on.

Previous Next