Jekyll2018-06-27T11:51:39+00:00http://www.jgyoung.ca/Jean-Gabriel YoungPersonal website for the research work and blog posts of Jean-Gabriel Young
Postdoctoral affiliation2018-06-26T00:00:00+00:002018-06-26T00:00:00+00:00http://www.jgyoung.ca/news/2018/06/26/lsa_cscs<p>This is not a
<a href="https://www.dynamica.phy.ulaval.ca/index.php?id=news&tx_ttnews%5Btt_news%5D=127&cHash=c75d462f29405c3c098aee6e01fd894b">well kept</a>
<a href="https://lsa.umich.edu/cscs/news-events/all-news/search-news/welcome-jean-gabriel-young-.html">secret</a>
<a href="https://www.jsmf.org/apply/fellowship/letters-of-intent.htm">anymore</a>, so I should post this here as well:
I’m joining the <a href="https://lsa.umich.edu/cscs/">Center for the Study of Complex System</a> at the University of Michigan as a Postdoctoral Fellow.
I’m looking forward to all the new exciting projects that will come out of this.</p>This is not a well kept secret anymore, so I should post this here as well: I’m joining the Center for the Study of Complex System at the University of Michigan as a Postdoctoral Fellow. I’m looking forward to all the new exciting projects that will come out of this.Netsci 20182018-04-16T00:00:00+00:002018-04-16T00:00:00+00:00http://www.jgyoung.ca/conferences/2018/04/16/netsci2018<p>I’ll be at NetSci 2018 (Paris) to present our recent preprint “<em><a href="https://arxiv.org/abs/1803.09191">Network archaeology: phase transition in the recoverability of network history</a></em>”.
The <a href="https://www.netsci2018.com/conference-program">program</a> is already out: I’m part of the <strong>Theory-I</strong> parallel session, on day 1 of the main event.</p>I’ll be at NetSci 2018 (Paris) to present our recent preprint “Network archaeology: phase transition in the recoverability of network history”. The program is already out: I’m part of the Theory-I parallel session, on day 1 of the main event.A few additions to the publication page2018-04-11T00:00:00+00:002018-04-11T00:00:00+00:00http://www.jgyoung.ca/publications/2018/04/11/a_few_papers<p>Two papers on which I appear as a co-author have recently been published in <strong>Physical Review E</strong>!</p>
<p>First, a deep cut by Guillaume St-Onge: <a href="https://doi.org/10.1103/PhysRevE.97.022305">The paper</a> tackles SIS dynamics on time-varying networks with fixed degree sequences. By changing the relative time-scale of the epidemics and of the network’s evolution, we are able to effectively interpolate between quenched and annealed formalisms of SIS dynamics on networks, thereby unifying <em>many</em> theoretical frameworks in one. My favourite result comes towards the end of the paper, where we show that the endemic phase can be <em>heterogeneous</em> near its onset: If you look at high degree nodes, then you’ll find that the disease’s prevalence scales faster with the network’s size than if you had inspected low degree nodes.</p>
<p>Second, a fun <a href="https://doi.org/10.1103/PhysRevE.97.032302">numerical paper</a> with <a href="http://edwardlaurence.me/">Edward Laurence</a>, <a href="https://sites.google.com/site/svmelnik/">Sergey Melnik</a> and <a href="https://www.dynamica.phy.ulaval.ca/index.php?id=contact">Louis J. Dubé</a>, where we show how to <em>exactly</em> solve cascade dynamics on small networks. By exact, I mean that we show how to calculate the probability of every single outcome, with arbitrary precision. Our algorithm is of course of exponential complexity, because there are exponentially many outcome in the first place; but there’s a few trick involved in reaching such a “simple” algorithm. <a href="https://arxiv.org/abs/1802.08849">Exact algorithms</a> are on the rise again!</p>
<p>I should also mention that I have recently uploaded a preprint <a href="https://arxiv.org/abs/1803.09191">to the arXiv</a>. This joint work with the extended <a href="https://www.dynamica.phy.ulaval.ca/">Dynamica</a> family grew out of a workshop held in 2016.
Our goal was to infer the past states of a network given its current structure. The paper characterizes the problem thoroughly, from the point of view of statistical inference. I’ll present this work at <a href="https://www.netsci2018.com/">NetSci 2018</a> during the first parallel session. Comments are more than welcome!</p>Two papers on which I appear as a co-author have recently been published in Physical Review E! First, a deep cut by Guillaume St-Onge: The paper tackles SIS dynamics on time-varying networks with fixed degree sequences. By changing the relative time-scale of the epidemics and of the network’s evolution, we are able to effectively interpolate between quenched and annealed formalisms of SIS dynamics on networks, thereby unifying many theoretical frameworks in one. My favourite result comes towards the end of the paper, where we show that the endemic phase can be heterogeneous near its onset: If you look at high degree nodes, then you’ll find that the disease’s prevalence scales faster with the network’s size than if you had inspected low degree nodes. Second, a fun numerical paper with Edward Laurence, Sergey Melnik and Louis J. Dubé, where we show how to exactly solve cascade dynamics on small networks. By exact, I mean that we show how to calculate the probability of every single outcome, with arbitrary precision. Our algorithm is of course of exponential complexity, because there are exponentially many outcome in the first place; but there’s a few trick involved in reaching such a “simple” algorithm. Exact algorithms are on the rise again! I should also mention that I have recently uploaded a preprint to the arXiv. This joint work with the extended Dynamica family grew out of a workshop held in 2016. Our goal was to infer the past states of a network given its current structure. The paper characterizes the problem thoroughly, from the point of view of statistical inference. I’ll present this work at NetSci 2018 during the first parallel session. Comments are more than welcome!JSMF Postdoctoral Fellowship2017-12-04T00:00:00+00:002017-12-04T00:00:00+00:00http://www.jgyoung.ca/news/2017/12/04/jsmf<p>While it’s been announced some times ago at <a href="https://www.lefil.ulaval.ca/lunivers-complexe-de-jean-gabriel-young/">Université Laval</a>, I haven’t posted this news here yet. So here goes: I’m beyond <em>thrilled</em> to announce that I have been selected by the <a href="https://www.jsmf.org/programs/cs/">James S. McDonnell Foundation</a> for their Postdoctoral Fellowship Awards in “<em>Understanding Dynamic and Multi-scale Systems</em>”.</p>
<p>I have not yet decided where I will be pursuing my postdoctoral studies, but one thing is certain: the next few years are going to be both intellectually stimulating and fun.</p>While it’s been announced some times ago at Université Laval, I haven’t posted this news here yet. So here goes: I’m beyond thrilled to announce that I have been selected by the James S. McDonnell Foundation for their Postdoctoral Fellowship Awards in “Understanding Dynamic and Multi-scale Systems”. I have not yet decided where I will be pursuing my postdoctoral studies, but one thing is certain: the next few years are going to be both intellectually stimulating and fun.The shape of randomness2017-09-25T00:00:00+00:002017-09-25T00:00:00+00:00http://www.jgyoung.ca/publications/2017/09/25/scm-published<p>I’m happy to announce that the results of a fun project with <a href="http://apatania.altervista.org/">Alice Patania</a>, <a href="https://lordgrilo.github.io/">Giovanni Petri</a> and Francesco Vaccarino are now available in a short paper in <a href="https://doi.org/10.1103/PhysRevE.96.032312">Physical Review E</a>.
Special thanks to the <a href="http://yrncs.cssociety.org/">YRNCS</a> who funded this collaboration through their <a href="http://yrncs.cssociety.org/bridge-grants/">Bridge Grant initiative</a>.</p>
<p>Our paper, titled “<em>Construction of and efficient sampling from the simplicial configuration model</em>”, builds on <a href="https://arxiv.org/abs/1602.04110">recent work</a> by Owen T. Courtney and Ginestra Bianconi.
More specifically, we propose an efficient and provably correct MCMC algorithm for the maximally random ensemble of <a href="https://en.wikipedia.org/wiki/Simplicial_complex">simplicial complexes</a> with given degree and dimension sequences.
This algorithm comes in handy when we use simplicial complexes as an abstraction for the structure of complex systems (as has been done <a href="https://link.springer.com/article/10.1140/epjds/s13688-017-0104-x">more and more often recently</a>).
By randomizing a simplicial complex that encodes the structure of some system <em>X</em>, we get to tell what connection patterns in <em>X</em> are explained by simple local properties (degrees and dimensions), and what patterns are surprising.
In our paper, we this method to show that the <a href="https://en.wikipedia.org/wiki/Homology_(mathematics)">homology groups</a> of a few real-world systems are decidedly not random, and we suggest some mechanisms that could explain these differences.
But one could use the method to investigate different property; this will be dictated by application!</p>
<p>The next step will be to develop correlated variant of the ensemble (think of the degree-correlated CM) and to delve into combinatorial results (e.g., rigorously prove the traversability of the associated simplicial complex space).</p>
<!--more-->
<p><strong>Added on November 18th, 2017:</strong> <a href="https://www.facebook.com/KRStories/">Kendra Redmond</a> wrote <a href="http://physicsbuzz.physicscentral.com/2017/10/the-shape-of-randomness.html">a nice blog post</a> about our paper, over at the APS’s Physics Central. Check it out!</p>I’m happy to announce that the results of a fun project with Alice Patania, Giovanni Petri and Francesco Vaccarino are now available in a short paper in Physical Review E. Special thanks to the YRNCS who funded this collaboration through their Bridge Grant initiative. Our paper, titled “Construction of and efficient sampling from the simplicial configuration model”, builds on recent work by Owen T. Courtney and Ginestra Bianconi. More specifically, we propose an efficient and provably correct MCMC algorithm for the maximally random ensemble of simplicial complexes with given degree and dimension sequences. This algorithm comes in handy when we use simplicial complexes as an abstraction for the structure of complex systems (as has been done more and more often recently). By randomizing a simplicial complex that encodes the structure of some system X, we get to tell what connection patterns in X are explained by simple local properties (degrees and dimensions), and what patterns are surprising. In our paper, we this method to show that the homology groups of a few real-world systems are decidedly not random, and we suggest some mechanisms that could explain these differences. But one could use the method to investigate different property; this will be dictated by application! The next step will be to develop correlated variant of the ensemble (think of the degree-correlated CM) and to delve into combinatorial results (e.g., rigorously prove the traversability of the associated simplicial complex space).Voter model on the adaptive SBM2017-08-10T00:00:00+00:002017-08-10T00:00:00+00:00http://www.jgyoung.ca/publications/2017/08/10/sbvm-published<p>In January 2016, <a href="http://laurenthebertdufresne.github.io/">Laurent Hébert-Dufrense</a>, <a href="http://antoineallard.github.io/">Antoine Allard</a>, <a href="https://scholar.google.com/citations?user=FxU9cG0AAAAJ">Pierre-André Noël</a>, <a href="http://ericlibby.github.io/">Eric Libby</a> and I got together for a <a href="https://www.santafe.edu/">SFI</a> working group, to investigate competitor dynamics on networks.
Our starting point was the connection between <a href="https://en.wikipedia.org/wiki/Voter_model">voter models</a> and <a href="https://en.wikipedia.org/wiki/Moran_process">the Moran process</a>, and, more generally, questions about the political arena and biology.
Our answers are now available in a <em>Scientific Reports</em> paper titled “<a href="http://dx.doi.org/10.1038/s41598-017-07621-x"><em>Strategic tradeoffs in competitor dynamics on adaptive networks</em></a>”.</p>
<p>In the paper, we introduce a voter model on the adaptive SBM (the structure of the network changes depending on who’s claiming what resources).
It turns out that the model can be mapped to a well-known evolutionary game theory problem.
<strong>The upshot?</strong> This gives us a game-theoretical perspective on network structure.
In turn, this allows us to conclude that, for example, sustaining echo chambers is not a robust and viable strategy.
So intead of fostering closed communities, strive for open discourse across boundaries.</p>
<!--more-->
<p>Check the <a href="https://phys.org/news/2017-08-algae-political-strategy.html">Phys.org</a> article for a popular summary.</p>In January 2016, Laurent Hébert-Dufrense, Antoine Allard, Pierre-André Noël, Eric Libby and I got together for a SFI working group, to investigate competitor dynamics on networks. Our starting point was the connection between voter models and the Moran process, and, more generally, questions about the political arena and biology. Our answers are now available in a Scientific Reports paper titled “Strategic tradeoffs in competitor dynamics on adaptive networks”. In the paper, we introduce a voter model on the adaptive SBM (the structure of the network changes depending on who’s claiming what resources). It turns out that the model can be mapped to a well-known evolutionary game theory problem. The upshot? This gives us a game-theoretical perspective on network structure. In turn, this allows us to conclude that, for example, sustaining echo chambers is not a robust and viable strategy. So intead of fostering closed communities, strive for open discourse across boundaries.Detectability of the SBM in Phys. Rev. E and more!2017-06-19T00:00:00+00:002017-06-19T00:00:00+00:00http://www.jgyoung.ca/publications/conferences/2017/06/19/detectability-of-the-sbm-in-pr<p>So, two short news.</p>
<p>First, “<em>Finite size analysis of the detectability limit of the stochastic block model</em>” is now published in <a href="https://journals.aps.org/pre/abstract/10.1103/PhysRevE.95.062304">Phys. Rev. E</a>; I’ll upload an updated version to the arXiv soon (after <a href="http://netsci2017.net/">NetSci 2017</a>).
It is a long paper, so it was a long process.
I’m glad to see it through!
Our coolest result are, I believe, <strong>1.</strong> the symmetry group of the SBM and <strong>2.</strong> our approximation solution of the hypersurface equations.
The first tells us what transformations of the parameters maintain the difficulty of the detectability/recovery problem, while the second determines the surface of constant detectability in the parameter space.</p>
<p>Second news: I’ll be giving an extra talk at NetSci, during the satellite sessions <a href="http://complexdata.businesscatalyst.com/program.html"><strong>tomorrow</strong></a> (I’ve already announced this earlier <a href="https://twitter.com/_jgyou/status/874687304065011712">on Twitter</a>). I’ll present the <a href="https://arxiv.org/abs/1705.10298">Simplicial Configuration Model</a> and some recent related work done with <a href="http://apatania.altervista.org/">Alice Patania</a>, <a href="https://lordgrilo.github.io/">Giovanni Petri</a> and Francesco Vaccarino.
<!--more-->
The slides should be up on <a href="https://speakerdeck.com/jgyou/">speakerdeck</a> by the end of the talk!</p>So, two short news. First, “Finite size analysis of the detectability limit of the stochastic block model” is now published in Phys. Rev. E; I’ll upload an updated version to the arXiv soon (after NetSci 2017). It is a long paper, so it was a long process. I’m glad to see it through! Our coolest result are, I believe, 1. the symmetry group of the SBM and 2. our approximation solution of the hypersurface equations. The first tells us what transformations of the parameters maintain the difficulty of the detectability/recovery problem, while the second determines the surface of constant detectability in the parameter space. Second news: I’ll be giving an extra talk at NetSci, during the satellite sessions tomorrow (I’ve already announced this earlier on Twitter). I’ll present the Simplicial Configuration Model and some recent related work done with Alice Patania, Giovanni Petri and Francesco Vaccarino.Presenting at NetSci 20172017-03-25T00:00:00+00:002017-03-25T00:00:00+00:00http://www.jgyoung.ca/conferences/2017/03/25/netsci<p>My submitted talk “<em>Statistical mechanics of mesoscopic structure extraction</em>” has been accepted at <a href="http://netsci2017.net/">NetSci
2017</a>. I will present an unifying view—already explored by many, but not
quite complete—of
community
detection and mesoscopic structure in general, using the language of statistical mechanics.
This is joint work with members of <a href="http://www.dynamica.phy.ulaval.ca">my research group</a>, who will also present quite a few
talk
of their own, see <a href="http://www.dynamica.phy.ulaval.ca/index.php?id=conferences">this list of abstracts</a>.
<!--more-->
See you there!</p>My submitted talk “Statistical mechanics of mesoscopic structure extraction” has been accepted at NetSci 2017. I will present an unifying view—already explored by many, but not quite complete—of community detection and mesoscopic structure in general, using the language of statistical mechanics. This is joint work with members of my research group, who will also present quite a few talk of their own, see this list of abstracts.Jekyll rewrite2016-09-23T00:00:00+00:002016-09-23T00:00:00+00:00http://www.jgyoung.ca/miscellaneous/2016/09/23/an-experiment<p>I’m now managing my website’s content with <a href="https://jekyllrb.com">Jekyll</a>.
The rewrite took some time, but it is definitely worth it.
Jekyll is much more powerful, flexible and efficient than my previous static website generator.
Not that it was exactly hard to beat—up until now, I handled my website with a <a href="https://github.com/jg-you/science-static">buggy, incomplete, and partial Jekyll clone of mine</a>.
Back then, my goal was to learn python with a practical project, and it definitely helped. But why reinvent the wheel when you know how to drive? Out with <code class="highlighter-rouge">static-science</code>, ìn with <code class="highlighter-rouge">jekyll</code>.
Oh, and I also took this opportunity to reskin the site.
<!--more--></p>
<p>What do you think?</p>I’m now managing my website’s content with Jekyll. The rewrite took some time, but it is definitely worth it. Jekyll is much more powerful, flexible and efficient than my previous static website generator. Not that it was exactly hard to beat—up until now, I handled my website with a buggy, incomplete, and partial Jekyll clone of mine. Back then, my goal was to learn python with a practical project, and it definitely helped. But why reinvent the wheel when you know how to drive? Out with static-science, ìn with jekyll. Oh, and I also took this opportunity to reskin the site.New paper in Phys. Rev. E2016-09-15T00:00:00+00:002016-09-15T00:00:00+00:00http://www.jgyoung.ca/publications/2016/09/15/new-paper-in-phys-rev-e<p>My latest paper with <a href="http://laurenthebertdufresne.github.io/">L. Hébert-Dufresne</a>, <a href="http://antoineallard.info">A. Allard</a> and <a href="http://dynamica.phy.ulaval.ca/">L.J. Dubé</a> is now out in Physicial Review E.</p>
<p>In the paper, titled <a href="http://dx.doi.org/10.1103/PhysRevE.94.022317">“Growing networks of overlapping communities with internal structure”</a>, we come up with a natural and dynamical explanation of the Dunbar number, i.e., an upper bound on the number of connection that individuals can sustain in a social network.
We show how this number is related to the heterogeneity of the degree distribution of nodes <em>within</em> communities and investigates the consequences of this explanation for networks which have a growing and overlapping community structures.</p>My latest paper with L. Hébert-Dufresne, A. Allard and L.J. Dubé is now out in Physicial Review E. In the paper, titled “Growing networks of overlapping communities with internal structure”, we come up with a natural and dynamical explanation of the Dunbar number, i.e., an upper bound on the number of connection that individuals can sustain in a social network. We show how this number is related to the heterogeneity of the degree distribution of nodes within communities and investigates the consequences of this explanation for networks which have a growing and overlapping community structures.