Another great guest is visiting my group this week: Conrad Lee. Conrad has been writing consistently superb blog posts over at http://sociograph.blogspot.com for quite a while now (I highly recommend checking out his back catalog, which contains insightful analysis of issues related to community detection in complex networks and much more). And he’s interested in many of the same topics that I’ve worked on for years, so there should be lots of great discussions.
Tomorrow, he’ll be speaking on methods for validating community detection algorithms using meta-data in the talk: Are network communities good for nothing? Benchmarking algorithms with inference tasks. With abstract hinting at a very interesting talk (and containing wildlife simile):
While community detection algorithms proliferate like rabbits in the spring, relatively little work has gone into determining which methods work best. In many cases, we know only that a given method can partition Zachary’s Karate club – a problem which was solved over thirty years ago. Furthermore, the small literature concerned with benchmarking these algorithms focuses on synthetic data, leaving us with little evidence to support the claim that we can find meaningful communities in non-trivial, real-world social network data. We know so little about the performance of these algorithms because on the one hand we have a poor a priori intuition of how network communities are actually structured, and on the other hand we lack datasets that have a “ground truth” set of communities.
In this presentation, I argue that the quality of network communities can be evaluated by measuring how well they allow inference of missing information, such as certain node attributes and missing links. More concretely, good network communities should provide a machine learning model with informative features. I will discuss some conceptual and practical difficulties which came up when implementing a benchmark based on this premise using the Facebook100 dataset. Early results indicate that all tested methods have a bias for a particular scale, a finding which suggests that a scaling parameter is necessary. For example, modularity maximization and the Map Equation perform poorly, even when using the hierarchical versions of these methods. Their performance improved only when using their generalized formulations, which include a scaling parameter that alters the underlying objective function.
I highly recommend stopping by if you’re in the area! Time and place are listed here.
This week, we have another exciting guest, Bruno Gonçalves (twitter: @bgoncalves) will be visiting the lab Monday 24th of September, and Tuesday the 25th. Bruno has just moved to the university at Aix-Marseille University, from Alex Vespignani’s group at Northeastern and we’re excited to have him.
Bruno is giving a talk Monday at 11 – I highly recommend it:
Title: From Individual Activity to Collective Attention – Insights from Large Scale Social Network Analysis
Abstract: Modern social systems such as Twitter expose digital traces of social discourse with an unprecedented degree of resolution of individual behaviors. They offer an opportunity to investigate both individual and collective behavioral patterns and to disentangle the temporal, spatial and topical aspects of human activity.
A large survey of online exchanges or conversations on Twitter, collected across six months involving 1.7 million individuals is used to study how individuals manage their social relations. Two main features are observed:
1. Social interaction strength is highly dependent of the number of connections, corroborating Dunbar’s Social Brain theory. A simple model shows how limited individual capacity for social interaction is enough to qualitatively reproduce the features observed.
2. Users display extremely diverse activity levels that follow a broad tailed distribution. We construct an activity driven model that is capable of encoding the instantaneous time description of social network dynamics. Within this framework, highly dynamical networks can be described analytically, providing a powerful tool for the analysis of social phenomena occurring over time-varying networks.
Finally, we focus on Twitter activity surrounding American Idol voting as minimal and simplified version of complex societal phenomena such as political elections, and show that the volume of information available in online systems permits the real time gathering of quantitative indicators anticipating the future unfolding of opinion formation events.
Time & place:
Monday 24 September at 11:00-12:00
Building 305, Seminar room 053
This week, we’ll have two exciting visitors. Sebastian Ahnert will be visiting on Wednesday, speaking on Compressible components reveal network architectures.
On Thursday we’ll change gears and discuss mobility, modelling, and sensing. For that we’re lucky to have Cecilia Mascolo and her postdoc Neal Lathia visiting, speaking on How can sensor data be used to infer how people feel?
If you’re anywhere near Copenhagen, I highly recommend both talks!
So catching up on recent developments, a big item from last month is that I made the cover of the local university paper DTU avisen – with nice photos and everything.
The best version of the article is only available in Danish [find it here]. Due to a bit of last minute editorial changes, a bunch of errors made their way into the printed version [find it here DK link, GB link (pdf), GB link (html)]
Just a quick post to note that I’m in Boston and will be around for much of the summer (the rest of July and the first part of August). I arrived a bit over a week ago, but have been too busy to update the blog – I guess better late than never. In fact I have a substantial blog-backlog, so expect more activity on the blog over the next few weeks.
I’ll be spending my time at Center for Complex Network Research and the LazerLab. Do send me an email if you’re around and would like to meet up!
Just back from a great NetSci conference at Northwestern University. Aside from being an excellent meeting with plenty of interesting talks and opportunities to meet with old friends, the conference was a reminder that NetSci 2013 is approaching rapidly.
At the conference, Petter Holme and I managed to recruit a great set of new committee members to help organize next year’s conference. In addition to Petter and myself, the organizing team now counts
- Katharina Zweig as the technical chair, responsible for submissions.
- Jan O. Haerter is the chair of registration.
- Isabel Meirelles has graciously promised to do a redesign of the NetSci visual identity.
- Philipp Hövel will be the chair of satellites (so let him know if you’re interested in running a satellite event).
- Bruno Gonçalves is going to be the web chair, responsible for the content of the netsci website.
- In Copenhagen Joachim Mathiesen is organizing all things local.
- And finally, Alan Mislove runs the NetSci web-infrastructure with a steady hand.
With this amazing team, help from the Network Science Society, and a superb venue (see below), I’m sure that NetSci 2013 is well under way. But we could always use more help, so let me know if you have any ideas for how you can help make NetSci 2013 the best ever. We’re also always looking for volunteers to help us run the conference itself.
This week, my collaborator/friend from Barabasilab/award winning physicist, Yong-Yeol Ahn (better know as YY) is visiting the Center for Social Data here at DTU. If you’re anywhere near Copenhagen, I highly recommend you stop by to see his talk!
Time: Thursday, May 24th, 13:00 [details here]
Title: Community structure and flocking of memes in social networks
Abstract: Spreading processes on networks (e.g. epidemic outbreak and information spreading) has been one of the most fundamental topics in network science. Information spreading in social networks has often been described by epidemic spreading models but recent studies demonstrated that some contagions (memes) exhibit fundamentally different pattern, where multiple exposure significantly enhances the transmission probability of the contagion. The co-operativity of a contagion makes the spreading process more sensitive to clustered network structure. Here we investigate the relationship between network communities and spreading of hashtags in a Twitter network.
And if we’re lucky, maybe we can even talk him into speaking a bit about his interesting work on food pairings and molecular gastronomy!
This week I’m visiting the Max Planck Institute for the Physics of Complex Systems in Dresden, participating in the workshop Mathematical Physics of Complex Networks: From Graph Theory to Biological Physics.
It’s quite the honor to be invited to speak at a conference full of real physicists & bona fide graph theorists (although it’s putting my softening brain hard at work: there seems to be a lot more analytical results here than I’m used to these days).
This week I’m heading out to CompleNet 2012 in Melbourne Florida. It looks like a great conference with Laszlo Barabasi, Robert Bonneau & Sinan Aral headlining.
The conference will also feature an unreasonably high level of activity by yours truly with duties including an invited talk (8:40 on March 8th), chairing a technical session (on network metrics and models, 10:20 on March 7th), as well as a brief talk at the opening of the art exhibition on The Art of Networks at the local Foosaner Art Museum, about the creation of the TwitterMood visualization.
Hope to see you there if you’re in or around the Sunshine State!
Renaud Lambiotte is visiting for a few days. He’s an exciting guy whose work focuses on the relation between dynamics, function and structure in complex systems, with a focus on neuronal and social networks (check his website for more details). He’s an associate professor in Mathematics at the University of Namur (Belgium).
If you’re in the Copenhagen area, I highly recommend going to his talk this friday. Here are the details.
Title: Random Walks on Networks: Dynamics and Teleportation
Abstract: In this talk, I will focus on two problems related to random walks on networks. First, I will focus on random teleportation, which is a necessary evil for ranking and clustering directed networks based on random walks. Teleportation enables ergodic solutions, but the solutions must necessarily depend on the exact implementation and parametrization of the teleportation. For example, in the commonly used PageRank algorithm, the teleportation rate must trade off a heavily biased solution with a uniform solution. Here we show that teleportation to links rather than nodes enables a much smoother trade-off and effectively more robust results, and discuss the effect of teleportation on clustering. In the second part of my talk, I will focus on random walks on temporal networks, i.e. networks evolving in time. In particular, I will examine the effects of inter-event statistics on the dynamics of edges, and apply the concept of a generalized master equation to the study of continuous-time random walks on networks.
Time & place:
- Friday 2 March 2012 at 14:00
- Technical University of Denmark
- Seminar room 053, Building 305
Everyone is welcome!