Accions del document
DELIS, scientists connected through research on networks and computing
LSI Department joined the most prominent dots in the computing and networks research during the DELIS workshop. National and international researchers put in common all their knowledge to spawn a bigger knowledge network
During four days we were able to enjoy the company of great professors like Joan Feigenbaum from Yale University or Tanya Berger-Wolf from the University of Illinois. We, as +LSI reporters, attended their presentations at DELIS workshop as well as other ones given by leading researchers from the LSI department and other universities. Of course there was a lot of learning in other settings: and we enjoyed lunches and a dinner full of scientific and worldly conversations.
On February 26th DELIS welcomed the audience and the adventure began. February 27th and 28th were intensive days full of presentations and discussions. It all ended on a final conclusion day, the 29th, and a sad goodbye.
One of the main goals of the event was promoting networking between researchers. It was not a surprise considering that we live in a planet that is getting unified and globalized thanks to networks. Computing networks were born some forty years ago. They have helped in creating a full body of theory about networks that is applied to very diverse fields. We explain in some details a couple of DELIS' lectures, each one with its own vision about networks and computing.
|
Theory of Networked Computation
by Joan Feigenbaum
|
Dynamic Social Networking for Zebras
by Tanya Berger-Wolf
|
|
The increasing prominence of the Internet during the last few years has created the necessity for theory of networked computation. The first question to be made is: What is going to be of computers during the next century? We don't have a theory yet, this is a challenge that we must face.
Economics became a key subject in professor Joan Feigenbaum's lecture. She talked about the problems related to the interdomain routes control on the internet, for example. These routes are based on local policies which are complex, non-coordinated and some times, private. Knowing the policies doesn't amount at all to have the problem solved; we need to look for a right algorithm that can represent such policies. This then brings us to some economic reasoning. When we analyse the routes on a computer communication network we can see that each person or company chooses the networks that suits him better usually in terms of cost. However, this can result in several problems. Oscillation is a typical one. Let's try to see what is it about. Let us suppose that we have two nodes, 1 and 2, which want to send some information to 3 through their neighbour, 2 and 1 respectively, because they think it is the better choice. Information will flow from 1 to 2 and then from 2 to 1, but it will never reach 3. Such a disaster should be solved by knowing each neighbour's actions. Unfortunately, knowing the neighbour's actions can also involve another problem. Let us imagine a new situation where at the beginning there is no oscillation because 1 is sending the information to 3 straight away, as well as sending it to 2. However, both 1 and 2 would prefer to use their neighbour to send the information because knowing that the other neighbour is part of a safe path will change their choice, generating a new oscillation situation. In this case, the mistake has been caused by working simultaneously. On top of that we can also have wheel situations where each neighbour sends the information to the following one and so on... Non-viable!
This is a great opportunity to do novel theoretical work that has practical impact. |
Tanya Berger-Wolf wanted to understand the creation better and evolution of dynamic social network models. He resorted to an animal group, zebras. She wanted to explore how zebras connect with each other within a group, and how the individual relationships start and evolve. Moreover, she wanted to represent this by means of a graph model that could include all the aspects of the dynamics of the group. Besides the interest of zebras as a test group for theoretical dynamic graph models of social networks, Tanya Berger-Wolf's endeavour had a second but not less important goal: to give insights to ecologists and conservationists in order to devise better protection policies for endangered species, for example, the grevy zebras, a seriously endangered African quadruped.
Tanya works with biologists from Princeton University. The field research with zebras is done in Africa, so she jokes about it saying that she had never thought that computing would bring her to such an exotic spot. You can see that she loves her job.
In order to get a better idea about the work of these two researchers you can have a look at the following documents: where you will find material from Joan Feigenbaum's speech
|