Hiding in plain sight

The concept of a “network” is deceptively simple and incredibly vital to life as we know it.

“Without networks, we don’t exist,” says Hiroki Sayama, associate professor in the Department of Systems Science and Industrial Engineering (SSIE). Sayama’s primary research focus is complex dynamical networks.

A network, or “complex system,” is a collection of components that are connected to, and interact with, each other.

It is vague by necessity, since everything from honeybees collecting nectar to global banking, from snakes hunting mice to human consciousness itself, is a network.

“If you think about it, that very ability to think is network-based,” Sayama says. “Billions of tiny neurons and synapses firing in a certain order give us consciousness. Without each piece and each connection, there is no ‘us.’”

Thinking the unthinkable

The complexities of real-world networks — hidden to most of us — are tantalizing to scientists. Understanding nonlinear connections between multiple components in a system, or discovering simultaneous interactions among multiple systems, demands research that is complicated and abstract.

“You have to think the unthinkable,” Sayama says.

For example, repaving a road could lower health insurance premiums. The two seem unconnected, but think about this (using that network in your head):

The repaving lowers wait times for drivers, which is good for traffic flow.

Then there can be a ripple effect.

Drivers who are not frustrated by slow traffic and distracted by potholes can concentrate on their driving. This leads to fewer car accidents, fewer insurance claims and, as a result, lower premiums.

That all seems logical, but here’s the rub: The repaving might have little or no effect on traffic flow, driver distraction, accident rates or insurance claims.

This is where studying networks and complex systems comes in. Perhaps traffic-light sequencing is an issue, or pedestrian traffic is high. Maybe the orientation of the road to the rising or setting sun has something to do with everything. Or the repaving project might have had an impact in an unexpected place like the welfare system, because it put people to work.

“Everything described is just the beginning of thinking about how this particular system interacts,” Sayama says, laughing. “This is such a holistic field of study because everything is connected, woven together; all variables have a say in how the world is working.”

Also, those who study networks are interested in improving them at a larger scale than, for instance, making changes to a single road in hopes of achieving a single result.

“With a holistic approach, we are thinking about how to improve roads in an entire city,” Sayama says. “There can be surface results like reduced wait times in traffic, but ‘real’ results might not be obvious. Some positive impacts can be small but affect many people, so there is a large impact even though it isn’t really felt. The entire system is better, but it is difficult to quantify.”

Even with benefits, vague results make it hard to sell officials on further study.

“It is so important to adopt this kind of thinking, because if policymakers and administrators embrace it, they will improve everything instead of going after small problem after small problem,” Sayama says.

Networks cross many disciplines

How do you teach something described as “unthinkable”?

At Binghamton University, there is a growing network of people and resources to teach and study networks and complex systems.

In the classroom, students can earn an Advanced Graduate Certificate in Complex Systems Science and Engineering — the program is directed by Sayama — within the Thomas J. Watson School of Engineering and Applied Science. Mathematical modeling, simulation, nonlinear statistics and other techniques are taught and used to study how things connect in unique ways, as well as how stimuli affect large systems.

Sayama is also the director of the Center for Collective Dynamics of Complex Systems (CoCo). CoCo, which became an official organized research center in July 2015, brings together an interdisciplinary group of professors and students, along with outside experts, to study complex systems. The center holds a biweekly seminar series that has included topics such as “Termites as Models of Swarm Cognition,” “Greed, Speed, and Deception in the Evolution of Drug Resistance in Malaria” and “Estimating Economic Impact of Online Product Reviews.”

“We created CoCo and the certificate program to make people aware that there are tons of networks and complex systems around us. If you become aware of their presence, interacting with the world and the ability to make changes to society will be different,” Sayama says.

Understanding terrorists

Perhaps the most topical network research at Binghamton was presented at a CoCo seminar last spring.

Binghamton PhD candidate Salih Tutun studied over 140,000 terrorist attacks between 1970 and 2014 with the help of Mohammad Khasawneh, professor and SSIE department chair.

The results of the research were presented in a paper titled “Understanding Patterns and Relations of the Terrorist Attacks to Prevent Future Threats.” Co-authors were Assistant Professor Chun-An Chou from SSIE and PhD candidate Sina Khanmohammadi. The work was presented in May at the 2016 Industrial and Systems Engineering Research Conference in California.

A framework was developed to calculate the relationship between certain aspects of terrorist attacks, such as time of the attack and weapons used.

That same framework can then be extrapolated to predict characteristics of future terrorist attacks — with more than 90 percent accuracy, according to the study. The research says that terrorists tend to emulate the behavior of other terrorist groups and learn from mistakes and successes.

“[Terrorists] are learning, but they don’t know they are learning. We need to understand their patterns. Our framework works to define which metrics are important,” Tutun says. “For example, is there a relationship between the Paris attacks [the mass shootings and suicide bombings at cafés, a rock concert and a soccer match in November 2015] and the 9/11 attacks? If there is a relationship, we’re making a network. Maybe one attack in the past and another attack have a big relationship. We try to extract this information.”

Previous studies focused on understanding the behavior of individual terrorists rather than studying attacks in relation to one another. Tutun believes policymakers can use the approach for detection of terrorist activity and take precautions to prevent attacks.

“Predicting terrorist events is a dream, but protecting an area by using patterns is a reality. If you know the patterns, you can reduce the risks,” Tutun says.

The research also reflects the holistic approach of studying entire systems instead of focusing on individuals or small groups.

“When you solve the problem in Baghdad, you solve the problem in Iraq. When you solve the problem in Iraq, you solve the problem in the Middle East. When you solve the problem in the Middle East, you solve the problem in the world,” Tutun says.

Early lessons in networks

The work with networks and complex systems also goes beyond campus.

Sayama co-led a group of more than 30 network science researchers, educators, teachers and students who set up a framework of seven concepts for teaching elementary, middle and high school students about networks. The group developed a planning booklet for teachers titled, “Network Literacy: Essential Concepts and Core Ideas.”

Sayama has also worked with the Maine-Endwell and Vestal school districts and at the NetSci High Summer Camp to help students understand networks. NetSci High is short for Network Science for the Next Generation and is funded by the National Science Foundation.

“In the education system, things are divided and specialized. Math, social studies, history, language, science are all broken up. But for complex systems, all disciplines are closely related and work together,” Sayama says. “This isn’t about finding a right or a wrong answer. Studying and testing theories about networks is about integration, improvement and understanding.”

Studying and testing take some thinking, too. If you think about it, there’s a network that can do that. n

John Brhel contributed to this story.