‘We Give Meaning to Ones and Zeros’
Can we use computer models to predict human behavior? And why do bad news travel so much quicker than good news? Florian Guckelsberger and Lars Mensel sat down with Jure Leskovic to discuss.
The European: There is a famous movie scene in “The Matrix” where one character stares at data and can see the people behind the codes. How much does data tell you?
Leskovic: Oh, data tells you a lot. You just have to listen to it.
The European: You have looked at a lot of codified data in social networks. What has been the most surprising discovery?
Leskovic: Good question. On many occasions, when you go into data, you do it with some intuition. Most of the times you have to change this intuition because the data tells you that had a wrong metaphor in your mind. The surprising events, however, are when you find something and you get a bit scared, scared in the sense that you discover something very different from what the scientific community has been thinking so far. Then you ask yourself if you have a bug. On one occasion we found that when a network becomes bigger and bigger its degree of separation begins to shrink - you are connected to others through a smaller number of intermediate steps. You have a network that grows but a degree of separation that decreases. Isn’t it very counterintuitive to have an object that grows but gets smaller?
The European: So you generate answers by just looking at data?
Leskovic: Yes, in this case it was an empirical fact. It was unclear what was going on and then we found out by looking at it. We were scared at first but then discovered the background.
The European: From the viewpoint of philosophy of science, can you find out anything about the human being by only looking at data? Is the truth rooted in codes?
Leskovic: Isn’t this the whole premise of science and natural sciences in particular, this materialistic view? You want to observe and quantify phenomena. You goal is to identify new facts and their causes. The important thing is that you are able to repeat the thing.
I am not a philosopher but from an empirical point of view, the truth is the thing that always repeats itself. It is something that anyone can repeat over and over again and it always happens the same way. If you think of what we have on the web and social media, the truth is what people are saying. Think of crowdsourcing for instance. If you ask how good a restaurant is online and receive answers from the internet community, that is the truth.
The European: You explained how you are able to identify the most vulnerable spots of any given network. What is the political dimension of this observation?
Leskovic: Any kind of science can be used for the good or the bad. The same is true for networks and computer sciences. This is why the American Defense Department is so interested in our research.
The European: Does the American Defense Department fund your research?
Leskovic: Yes, they fund a lot. But their angle has changed. Until a few of years ago, they were interested in understanding terrorist networks. After 9/11 it was all about understanding their structures. It was about knowing where they are keeping nodes and connections, and how you can efficiently disrupt them. This is how they thought about it until recently - but when they talk about network research today, they want to understand population dynamics, How can we understand the uprisings in the Middle East? How can we facilitate the new social media? At least in the mindset of those funding agencies it went from “how do we destroy networks?” to “how do we know that the truth comes out?” People are able to organize themselves.