APS News | People and History

Albert-László Barabási, Network Scientist, Wants Physicists to Connect with Wider Audiences

An interview with the recipient of the 2023 Lilienfeld Prize.

Published Oct 13, 2022
Albert-László Barabási, winner of the 2022 Lilienfeld Prize for his work in — and communication of — network science.
Hamu és Gyémánt / Lábady István

What do the internet, neurons in the brain, and cellular metabolism have in common?

According to Albert-László Barabási, all are complex networks with hidden patterns. Barabási, a physicist at Northeastern University and leader in network science, is the recipient of the 2023 Julius Edgar Lilienfeld Prize, which awards physicists who’ve made “outstanding contributions” to the field and communicate with diverse audiences.

“Network science has [come] of age,” says Barabási.

After Barabási earned his doctorate in statistical physics in 1994 from Boston University, he took a postdoctoral role in New York City. It was there, over the winter holidays, that he picked up a book on problems in computer science and first learned about graphs and networks. “How many networks must be supporting this interesting, fabulous, complex city?” he thought. Those networks, he realized, lacked “a theory of their own.”

A few years later, Barabási was running a materials science laboratory at the University of Notre Dame and had just received a federal grant to study quantum dots. Then, in 1999, he and co-author Réka Albert published a paper in the journal Science asserting that an enormous range of networks were “scale-free,” meaning they followed mathematical rules called power laws.

The paper changed everything. Barabási told his lab members, “I have zero interest from now on in materials science. I want to use all my resources and energy to focus on networks.” He sought to redirect his federal grant toward network research, to no avail; the funding was revoked. Still, he dove into networks and never looked back.

Barabási’s work helped fuel modern network science. It has also spurred debate, as some scientists dispute the ubiquity of scale-free networks in the real world.

Today, Barabási directs Northeastern University’s Center for Complex Network Research and is a lecturer in Harvard Medical School’s Department of Medicine. He also co-leads a European Research Council project on network science. He’s authored three popular science books, and, most recently, co-authored, with Dashun Wang, “The Science of Science,” a book examining patterns of career success in science.

Barabási spoke with APS News about the changing landscape of network science and his views on why physicists must communicate their research widely.

This interview has been edited for length and clarity.

How has the field of network science changed over your career?

In 1994, I wrote a paper [on networks] that I could not publish anywhere. The feeling from the referees was, “Why do we care?” In 1999, when our first real network paper started to emerge, people started to become interested in networks. There was lots of puzzlement among even my physics colleagues: “What are we really trying to study? Neural networks or spin classes?”

I kept saying, “It’s about any kind of network out there. We're trying to find, like physicists do, universal organizing principles. What are the laws common across different networks?”

When we started studying networks, we were not entering with a solution to a random problem. We were defining a new problem. And we were building the community, one paper at a time.

In a talk you gave, you said that our society is becoming a “laboratory” because of a “huge amount of data” available to study. Does this reality make it increasingly urgent that physicists communicate about their research to wider audiences?

Absolutely. I would even go further. COVID showed us how important this whole line of data-based thinking — and the role of physicists — is. Before COVID, epidemic prediction was based on traditional statistical methods. After our paper in ‘99 came out, another physicist, Alessandro Vespignani, defined network epidemiology. Network epidemiology allowed Vespignani’s team and others to start predicting, even before COVID, that something bad [would] happen. That’s one reason why Vespignani's work shaped the White House's response to COVID.

This was possible because physicists realized that epidemic processes are fundamentally network-based processes. They’re [also] stochastic processes, so you need statistical mechanics to properly describe them. Physics has been able to make a big impact in the community.

The circumstances [have] forced us to communicate what we do, because many of these models have really impacted people's lives. They were governing vaccine distribution, shutdown, all those things. Many physicists were the driving force behind that.

What open research questions fascinate you?

[Some] of the most exciting work we're doing now is to focus on physical networks — networks like in the brain, like neurons. The links are physical objects; there are cables there that cannot cross each other. We realized a few years ago that much of network science has sidelined the question of the physicality of the links in systems, like [the] vascular system or the brain or metamaterials. Now, a big effort in my lab is to develop the mathematical foundations and the physics of how we describe physical networks.

How can physicists better communicate about their research to diverse audiences?

First thing, do it. I think that 50% of success is your willingness to step out from your own narrow community and talk to a wider audience. That doesn't need to be a book. It can be an article in a newspaper. It can be a talk to undergrads.

We as physicists [tend] to be driven by intellectual curiosity and the coolness of the results that we get. There's an inherent beauty in some of the results, and we're happy if a few people understand that. When you step out of that community, that transition is not easy. It is, even today, painful for me, when I talk about a new subject, to formulate the message of “Why does this matter?” But I spend time to answer that in a way that is accessible to non-physicists, as well.

In a talk about your book “The Formula: The Universal Laws of Success,” you described performance as focused on the person who’s performing, but success as focused on how others perceive a person’s performance. How could this distinction impact research and recognition in the scientific community?

It's very important for a scientist to understand the patterns that govern their performance, the reception of their performance, or their scientific impact.

Thanks to the fact that research papers published since 1900 have been digitized and processed in a way that is analyzable, we have started massive research projects to understand the quantitative patterns that describe how a successful scientific career emerges.

Those patterns are very robust, and ignoring those is like trying to build an airplane without knowing Newton's Laws. The way scientific success emerges is not a random process. It has reproducible features, and understanding those could help young researchers have their work accepted much more widely and much faster.

For that reason, we wrote “The Science of Science.” The goal was to make [these lessons] accessible to any scientist. What are the quantitative laws that govern scientific careers? How do we not overestimate some of the measurable things, like citations, when it comes time to reflect on somebody's career?

Rachel Crowell

Rachel Crowell is a science journalist based in Iowa.

writesrcrowell

Join your Society

If you embrace scientific discovery, truth and integrity, partnership, inclusion, and lifelong curiosity, this is your professional home.