Episode 10: What's the rumpus about Networks 2021?
Updated: Jun 14, 2021
What's the rumpus 😊
I'm Asaf Shapira and this is NETfrix - The Network Science Podcast.
In this episode we'll talk with Dr. Ann McCranie, the General Organizer of the conference, to learn what's so special about it and how to make the most out of it. We've managed to get her to spill the beans about soon-to-be-announced plenary talks and though she's impartial as chief organizer, she admitted to having a favorite presentation title. Care to guess what it is? In addition, to give us a taste of what's coming, we interviewed Prof. Baruch Barzel,
a renowned complex systems researcher, to talk about his 9, yes, nine presentations in the conference. But as NETfrix fans know – we judge by quality, not quantity, and luckily for us, Prof. Barzel got them both.
He'll unfold to us the super-interesting tale of his ongoing crusade to find universality in network dynamics. And like every good story, it will have some twists and turns and some surprising findings, including real world applications on Facebook. No spoilers.
So, to shed more light on the conference, we brought out the big guns, meaning Dr. Ann McCranie, the General Organizer of the conference and Associate Director of the Indiana University Network Science Institute: In the beginning - there was INSNA, the International Network for Social Network Analysis. It was founded in 1977 to encourage social research of human networks It's no wonder then that the field of networks owes a great debt to the Social Sciences - carrying this field on their shoulders for decades.
Then in 2006 the NETSCI society was established which carried the concept of "Network Science" to center stage. NETSCI is more oriented towards the fields of physics, computer science and mathematics. The differences between the two societies are also reflected in the data. In this case, Hendrik Schawe AKA @RandomGraphs built a Twitter network of the followers of the networks2021 conference. After applying Community Detection on this network, we can identify an SNA community, probably of INSNA members and another community of probably NETSCI members and yet another community that seems to be related to North-Eastern University that, as Ann says, share both worlds.
2021 is the first year that these two societies are joining forces and the story of how it came to be, appears in the interview with Ann on the podcast. She mentioned that the motivation to create "One Conference to Rule Them All" has been around for at least for 6 years. So, now we've got a connected component but is it a Giant Connected Component? It turns out that it's one of the biggest conferences ever held, spreading over 17 different time zones. When it comes to the number of presentations, there's no doubt that it's indeed the biggest. There are about 950(!) presentations in the main event alone.
But besides making history, the big thing about this conference is the diversity.
I guess there isn't a scientific field left untouched by the presentations: From Covid19, through NLP, machine learning, political and social network analysis, organizational research and exotic topics like network analysis of religious movements in the Roman Empire, through Bitcoin research to the use of semantic networks to give economic forecasts. And have we mentioned Covid19?
The conference consists of several types of events:
Though there are many events, I tend to believe that the juicier parts seem to happen in the satellite events.
Satellite events include presentations and discussions around particular topics in a more intimate way to allow for deeper and meaningful discussions. Here is a short list of the satellite events I plan to attend:
So, the first is "How social phenomena affect the structure of the network"
In regard to the presentations, I plan to attend those that deal with Community Detection, and especially the one by Leto Peel under the temporary name "Detectability of hierarchical communities in networks" which received a warm recommendation from Michele Coscia, a networks' guru and a guest in an earlier episode of NETfrix.
Neurology is another topic I have the hots for, so I plan to check out those presentations as well. But the presentation I'm the most curious about is Javier Boldo's about Football (AKA Soccer) Tracking Networks. The novel idea behind it is to build a network based on the ball's movement, player's position and more. What does it mean? I have no idea, that's why I plan to attend it.
So far so good, but the catch is the entry fee. The cost of registration ranges from tens of dollars per day to about 200 dollars for the entire conference. But thanks to the "Whova" app that will be used in the conference, we can access the content even a month after the conference and word has it that you can even play it in 1.5 speed for those of us who lack in patience.
And now, to our episode's keynote speaker 😊 Prof. Baruch Barzel runs the complex systems lab in Bar Ilan University. One could think that publishing papers with Barabasi would be the climax of his career but it seems for Baruch is on a nonlinear climb as can be seen in the itinerary of the upcoming Networks2021 conference: His lab is scheduled to present 9 presentations in the upcoming conference!
Baruch shared his lab's latest discoveries about the different dynamic patterns that run on the network's topology. We couldn’t cover all his lab's presentations, so we focused on the one with the best title: "Spatiotemporal signal propagation in complex networks".
By using different spreading models on the same network's topology, Baruch managed to show that different patterns evolve from the different models, although the topology stayed the same. To explain it, Baruch uses the "Waze" app metaphor: Though roads tend to be stable, as the network's topology is, the traffic, or the dynamic processes tend to change.
On the one hand, this discovery delivers bad news to those of us who seek to find universal laws in Network Science. On the other, Baruch shows that he found distinct dynamic patterns of spreading through the network, and he sums them up to 3 categories:
An intuitive pattern where the Hubs in the network help to spread the information throughout the network.
An unintuitive pattern where the Hubs buffer the information, causing the spread to bypass them.
A pattern where the spread is agnostic to the Degree of the nodes in the network.
A dynamic pattern where the Hubs play a role of buffers rather than spreaders sounds weird at first but here are a couple of possible explanations for it:
In NETfrix episode "6-Degrees of ihttps://bit.ly/NETfrix_6DEG_Blogntuition", we've shown how and why Hubs can be pictured as mountains in the network, making it difficult to us to travel through the network. Though I'm heavily biased, I recommend this episode to those who look for an intuitive explanation to these phenomena. In the episode we'll also see what it has to do with the mysterious Dunbar Number.
Lada Adamic's paper "Do Diffusion Protocols Govern Cascade Growth?" demonstrates on Facebook's data that different viral campaigns have different spreading strategies. A successful meme campaign will use the Hubs to spread fast in the network while a challenge campaign (like the "Ice water Bucket" challenge) will spread more slowly but will bypass the Hubs on its way. A possible explanation for this is that Hubs can find the time to pass a meme but challenges require more energy that they might can't spare because of the huge amount of links they need to maintain.
For those interested in following the field of networks in general and the conference in particular, you're welcome to follow @NetfrixP on Twitter.
Did you enjoy, and want to share? Have you suffered and you don't want to suffer alone?