Researchers investigate the veracity of ‘six degrees of separation’

Do you know someone who knows someone? We have all played this game, often to be amazed that despite the extreme scale of human society, random people can be linked through very small chains of acquaintances—typically, around six. Recently, a group of researchers from across the globe discovered that this magic of six degrees can be explained mathematically. The intriguing phenomenon, they show, is linked to another social experience we all know too well—the struggle of cost vs. benefit in establishing new social ties.

In 1967, a farmer in Omaha, Nebraska received a peculiar letter in his mailbox. The sender was Prof. Stanley Milgram, of Harvard University, and the intended recipient was one of his peers. “If you happen to know this person,” the message read, “please forward this letter to him.”

Of course, the chances of such a direct acquaintance across such a vast social and geographical distance—from Boston to Omaha—were extremely slim, and therefore, the letter further requested that if the recipient didn’t know the intended addressee, they should forward the letter to someone who might.

This letter was one of about 300 identical packages sent with similar instructions. The 300 independent letters began circulating across the United States in pursuit of a social pathway linking “Joe” from the farmlands of middle America with the academic hub of the East Coast. Not all letters made it through, but the ones that did recorded, for the first time experimentally, the familiar social paths—a friend of a friend of a friend—that connect American society.

Quite surprisingly, the paths were found to be extremely short. In a society of hundreds of millions of individuals, the experiment found that it only takes about six handshakes to bridge between two random people. Indeed, Milgram’s experiment confirmed what many of us sense intuitively, that we live in a small world, divided by a mere six degrees of separation.

As groundbreaking as it was, Milgram’s experiment was also shaky. For example, it did not count the letters that didn’t reach their final destination. Most letters never reached their destination in Boston. The few letters that actually did arrived through six steps on average. His findings, however, were reaffirmed in a series of more systematic studies: for example, the millions of users of Facebook are on average five to six clicks apart from one another. Similar distances were also measured across 24,000 email users, actor networks, scientific collaboration networks, the Microsoft Messenger network and many others. Six degrees kept coming up.

Hence, social networks of vastly different scale and context tend to feature extremely short pathways. And most importantly, they seem to universally favour the magic number of six. But why?

A recent paper published in Physical Review X by collaborators from Israel, Spain, Italy, Russia, Slovenia and Chile, shows that simple human behaviour—weighing the costs and benefits of social ties—may uncover the roots of this intriguing phenomenon.

Consider individuals in a social network. Naturally, they wish to gain prominence by navigating the network and seeking strategic ties. The objective is not simply to pursue a large number of connections, but to obtain the right connections—ones that place the individual in a central network position. For example, seeking a junction that bridges between many pathways, and hence funnels much of the flow of information in the network.

Of course, such centrality in the network, while offering extremely valuable social capital, does not come for free. Friendship has a cost. It requires constant maintenance.

As a result, the research shows, social networks, whether on or offline, are a dynamic beehive of individuals constantly playing the cost-benefit game, severing connections on the one hand, and establishing new ones on the other. It’s a constant buzz driven by the ambition for social centrality. At the end, when this tug-of-war reaches an equilibrium, all individuals have secured their position in the network, a position that best balances between their drive for prominence and their limited budget for new friendships.

“When we did the math,” says Prof. Baruch Barzel, one of the paper’s lead authors, “we discovered an amazing result: this process always ends with social paths centered around the number six. This is quite surprising. We need to understand that each individual in the network acts independently, without any knowledge or intention about the network as a whole. But still, this self-driven game shapes the structure of the entire network. It leads to the small world phenomenon, and to the recurring pattern of six degrees,” adds Prof. Barzel.

The short paths characterizing social networks are not merely a curiosity. They are a defining feature of the network’s behaviour. Our ability to spread information, ideas and fads that sweep through society is deeply ingrained in the fact that it only requires a few hops to link between seemingly unrelated individuals.

Of course, not only do ideas spread through social connections. Viruses and other pathogens use them, as well. The grave consequences of this social connectedness were witnessed firsthand with the rapid spread of the COVID pandemic that demonstrated to us all the power of six degrees. Indeed, within six infection cycles, a virus can cross the globe.

“But on the upside,” adds Prof. Barzel, “this collaboration is a great example of how six degrees can play in our favour. How else would a team from six countries around the world come together? This is truly six degrees in action!”

For more such insights, log into our website https://international-maths-challenge.com

Credit of the article given to Bar-Ilan University


New statistical tool to distinguish shared and unique features in data from different sources

When facing a daunting dataset, Principal Component Analysis (PCA), known as PCA, can help distill complexity by finding a few meaningful features that explain the most significant proportion of the data variance.

However, PCA comes with the underlying assumption that all data sources are homogeneous.

The growth in Internet of Things connectivity poses a challenge as the data collected by “clients,” like patients, connected vehicles, sensors, hospitals or cameras, are incredibly heterogeneous. As this increasing array of technologies from smartwatches to manufacturing tools collect monitoring data, a new analytical tool is needed to disentangle heterogeneous data and characterize what is shared and unique across increasingly complex data from multiple sources.

“Identifying meaningful commonalities among these devices poses a significant challenge. Despite extensive research, we found no existing method that can provably extract both interpretable and identifiable shared and unique features from different datasets,” said Raed Al Kontar, an assistant professor of industrial and operations engineering.

To tackle this challenge, the University of Michigan researchers Niaichen Shi and Raed Al Kontar developed a new “personalized PCA,” or PerPCA, method to decouple the shared and unique components from heterogeneous data. The results will be published in the Journal of Machine Learning Research.

“The personalized PCA method leverages low-rank representation learning techniques to accurately identify both shared and unique components with good statistical guarantees,” said Shi, first author of the paper and a doctoral student of industrial and operations engineering.

“As a simple method that can effectively identify shared and unique features, we envision personalized PCA will be helpful in fields including genetics, image signal processing, and even large language models.”

Further increasing its utility, the method can be implemented in a fully federated and distributed manner, meaning that learning can be distributed across different clients, and raw data does not need to be shared; only the shared (and not unique) features are communicated across the clients.

“This can enhance data privacy and save communication and storage costs,” said Al Kontar.

With personalized PCA, different clients can collaboratively build strong statistical models despite the considerable differences in their data. The extracted shared and unique features encode rich information for downstream analytics, including clustering, classification, or anomaly detection.

The researchers demonstrated the method’s capabilities by effectively extracting key topics from 13 different data sets of U.S. presidential debate transcriptions from 1960 to 2020. They were able to discern shared and unique debate topics and keywords.

Personalized PCA leverages linear features that are readily interpretable by practitioners, further enhancing its use in new applications.

For more such insights, log into our website https://international-maths-challenge.com

Credit of the article to be given Patricia DeLacey, University of Michigan College of Engineering