There are some brilliant epitaphs on these tombstones. I have seen a few of them before (such as "I Told You I Was Sick", which is a favorite), but there are a lot of great ones I hadn't yet seen.
My name is Mason Porter. I am a Professor in the Department of Mathematics at UCLA. Previously I was Professor of Nonlinear and Complex Systems in the Mathematical Institute at University of Oxford. I was also a Tutorial Fellow of Somerville College.
Sunday, December 27, 2020
RIP Phil Niekro (1939–2020)
The great knuckleballer and Hall of Fame pitcher Phil Niekro died in his sleep last night.
Yet another great baseball player has left us this year.
Yet another great baseball player has left us this year.
Friday, December 18, 2020
Gollum's Cover of "Nothing Compares 2 U"
This deepfake of Gollum covering the song "Nothing Compares 2 U" is wrong on many, many levels. It's also glorious.
Witness the power of deepfakes! If I ever again speak at a workshop about deepfakes, I am whipping out this video.
(This cover needs to be used as a Ride chaser.)
(Tip of the cap to Jennifer Ouellette.)
Witness the power of deepfakes! If I ever again speak at a workshop about deepfakes, I am whipping out this video.
(This cover needs to be used as a Ride chaser.)
(Tip of the cap to Jennifer Ouellette.)
RIP Peter Neumann (1940–2020)
This morning, I saw the terrible news that my former Oxford colleague Peter Neumann died today, supposedly due to complications from COVID-19 (although I can't verify if that is the known cause of death). Here is is Peter's Wikipedia entry.
I had the chance to interact with Peter quite a few times at Oxford. When I knew him, he was concentrating on mathematics history, and I contacted him with a query about some mathematical linguistics (concerning, for example, the use of "topological" only versus both "geometric" and "geometrical"). I knew that Peter was the right person to approach when I wanted to dive deeper than what I was able to do via Google. Here is my 2013 blog entry about that.
Peter was one of the people in Oxford who was always so kind and supportive of me (and, from what I can tell, of many others as well). When I got some pointed negative teaching evaluations, Peter invited me to The Queen's College to dinner to chat about things (and to give gentle advice and so on). But if I am getting rewarded with a dinner like that for messing up, it does create strange incentives. My students in my courses now are certainly benefiting from advice Peter gave me over the years. He also supported me in various fights on committees (we were on Teaching Committee together) and particularly in efforts that I made against "We've always done it this way." comments to try to change things in Oxford's Mathematical Institute.
(Tip of the cap to Card Colm Mulcahy.)
I had the chance to interact with Peter quite a few times at Oxford. When I knew him, he was concentrating on mathematics history, and I contacted him with a query about some mathematical linguistics (concerning, for example, the use of "topological" only versus both "geometric" and "geometrical"). I knew that Peter was the right person to approach when I wanted to dive deeper than what I was able to do via Google. Here is my 2013 blog entry about that.
Peter was one of the people in Oxford who was always so kind and supportive of me (and, from what I can tell, of many others as well). When I got some pointed negative teaching evaluations, Peter invited me to The Queen's College to dinner to chat about things (and to give gentle advice and so on). But if I am getting rewarded with a dinner like that for messing up, it does create strange incentives. My students in my courses now are certainly benefiting from advice Peter gave me over the years. He also supported me in various fights on committees (we were on Teaching Committee together) and particularly in efforts that I made against "We've always done it this way." comments to try to change things in Oxford's Mathematical Institute.
(Tip of the cap to Card Colm Mulcahy.)
Wednesday, December 16, 2020
"Inference of Edge Correlations in Multilayer Networks"
A new paper of mine came out in final form today. Here are some details.
Title: Inference of Edge Correlations in Multilayer Networks
Authors: A. Roxana Pamfil, Sam D. Howison, and Mason A. Porter
Abstract: Many recent developments in network analysis have focused on multilayer networks, which one can use toencode time-dependent interactions, multiple types of interactions, and other complications that arise in complexsystems. Like their monolayer counterparts, multilayer networks in applications often have mesoscale features,such as community structure. A prominent approach for inferring such structures is the employment of multilayerstochastic block models (SBMs). A common (but potentially inadequate) assumption of these models is thesampling of edges in different layers independently, conditioned on the community labels of the nodes. In thispaper, we relax this assumption of independence by incorporating edge correlations into an SBM-like model. Wederive maximum-likelihood estimates of the key parameters of our model, and we propose a measure of layercorrelation that reflects the similarity between the connectivity patterns in different layers. Finally, we explainhow to use correlated models for edge “prediction” (i.e., inference) in multilayer networks. By incorporating edgecorrelations, we find that prediction accuracy improves both in synthetic networks and in a temporal network ofshoppers who are connected to previously purchased grocery products.
Title: Inference of Edge Correlations in Multilayer Networks
Authors: A. Roxana Pamfil, Sam D. Howison, and Mason A. Porter
Abstract: Many recent developments in network analysis have focused on multilayer networks, which one can use toencode time-dependent interactions, multiple types of interactions, and other complications that arise in complexsystems. Like their monolayer counterparts, multilayer networks in applications often have mesoscale features,such as community structure. A prominent approach for inferring such structures is the employment of multilayerstochastic block models (SBMs). A common (but potentially inadequate) assumption of these models is thesampling of edges in different layers independently, conditioned on the community labels of the nodes. In thispaper, we relax this assumption of independence by incorporating edge correlations into an SBM-like model. Wederive maximum-likelihood estimates of the key parameters of our model, and we propose a measure of layercorrelation that reflects the similarity between the connectivity patterns in different layers. Finally, we explainhow to use correlated models for edge “prediction” (i.e., inference) in multilayer networks. By incorporating edgecorrelations, we find that prediction accuracy improves both in synthetic networks and in a temporal network ofshoppers who are connected to previously purchased grocery products.
Sunday, December 13, 2020
The Cleveland Baseball Team
The Major League Baseball team that is situated in Cleveland will no longer be called the Indians. Finally! This is long past due. Hopefully, the Atlanta Braves will eventually follow suit.
Tales from the ArXiv: "A Nudge Framework"
A new paper that just appeared on arXiv this evening is called "Steering the aggregative behavior of noncooperative agents: a nudge framework". I come from a Jewish family, so I am very familiar with this strategy.
Tuesday, December 08, 2020
RIP Jean-Paul Revel (1931–2020)
Cellular biologist Jean-Paul Revel, who was Caltech's Dean of Students for several years (including during some of my undergraduate time), died a few days ago.
À bientôt, Dean Revel.
Update (12/09/20): I am aware of what "À bientôt" means. I used that line because Dean Revel ended all of his articles with it. (So, in case you were worrying about me, there is no need to worry!)
À bientôt, Dean Revel.
Update (12/09/20): I am aware of what "À bientôt" means. I used that line because Dean Revel ended all of his articles with it. (So, in case you were worrying about me, there is no need to worry!)
What Happens on Zoom Stays on Zoom
I'm not traveling these days, so I figure this is how this type of post needs to go nowadays.