Leto Peel

New blog post:

"Higher orders need higher standards"

skewed.de/lab/posts/hi...

I discuss our current work disentangling misconceptions around "higher-order" networks: arxiv.org/abs/2602.16937

Explainer thread for the paper here: bsky.app/profile/tiag...

Use whatever abstraction you find more useful or convenient!

We're not saying that hypergraphs are bad or useless.

We're just criticizing the unfounded claim that they are required to represent multivariate interactions.

Very happy to be part of this project. My take on it is, hypergraph or network can describe the same systems and yield the same phenomena. The decision really comes down to if you want to work with matrices or higher-order tensors to explore these phenomena.

Higher-order links and broken standards The balance in the Force has been restored

I have also written for you a short semi-technical piece on #ComplexityThoughts about this:

open.substack.com/pub/manlius/...

Enjoy!

Maybe it’s time we stop calling graphs “pairwise”.

Multivariate coupling functions include HO interactions by design, hyperedges are optional.

Under the order-decomposability assumption, the same node-level dynamics can be written as infinitely many equivalent multilayer (“gauge”) models.

1/

New preprint w/ @tiago.skewed.de, @manlius.bsky.social & @thilogross.bsky.social!

Graph edges define the interaction domain — functional form is free

Hyperedges constrain by grouping nodes into subsets: special case, not a generalisation

Multilayer graphs represent any hypergraph — not vice versa

Out now: "Gender differences in collaboration and career progression in physics" we use collaboration #networks to identify differences between women and men in collaboration and career progression in physics.

Work with Mingrong She, Jan Bachmann and Fariba Karimi.

There's only one week left to register for our Network Science Summer School at Utrecht University! Learn how networks shape our social, economic, and biological worlds. Hands-on training in Python & R from leading experts. July 7-11, 2025. Register now at net-science.github.io #NetworkScience

And we're done!

Goodbye everyone, safe journeys home!

When I thought it'd be a party hosting #NetSci2025 in Maastricht I didn't think it would manifest so literally 🥂🎶🕺

Thanks to everyone who helped us put this whole event together. Thanks to everyone who joined us this week. Seeing so many happy network scientists was truly wonderful!

The lovely Netplace people and their wonderful energy and ideas created some great opportunities for connections and engagement. It was really nice to work with them!

Congratulations for these lovely contributions!

Time to storm the castle!

@yasasgari.bsky.social showing our work on temporal multiscale assortativity with @piratepeel.bsky.social at the algorithm and metrics session at #netsci2025

Congratulations! #NetSci2025 secret unlocked!

This is actually just retaliation to mess with all those network scientists who've been using their badges for "free" train rides.

#NetSci2025

I came at #NetSci2025 to present but I ended up being presented. Thanks @piratepeel.bsky.social

#netsci2025 is officially open! @piratepeel.bsky.social explaining the advanced AI used to make badges find their owners

This sounds awful!

Starting the SINM satellite at @netsciconf.bsky.social with an exciting talk by Daichi Kuroda!

No time to summarize 2.5h lecture in a few posts

Some of the most important messages can be found by reading #ComplexityThoughts at manlius.substack.com and by reading our paper w/ @piratepeel.bsky.social @tiago.skewed.de at www.nature.com/articles/s41...

If you are in Maastricht, say hello

/fin

It was a pleasure to lecture at the @netsciconf.bsky.social for an interested & interactive audience.

What's a model? Why do we model? Is our network a network? Can we restore broken and forgotten links?

I will post on #ComplexityThoughts about this lecture, which I have enjoyed to prepare.

🧵 1/

We saved that for Friday #spoilers

Funny, I had the same thought this morning...

The alarm you just heard at #NetSci2025 is the Dutch testing their emergency alert system that happens every first Monday of the month at 12 noon.

THE WORLD IS NOT ENDING

(yet)

The quiet before the storm.

Good morning #NetSci2025

Not long now...

@netsciconf.bsky.social

We're looking for a new colleague to join us in Maastricht as an Assistant/Associate Professor in statistical learning in our Data Analytics and Digitalisation department.

Deadline to apply: June 8th

vacancies.maastrichtuniversity.nl/job/Maastric...

📢 𝗦𝗽𝗲𝗮𝗸𝗲𝗿 𝗔𝗻𝗻𝗼𝘂𝗻𝗰𝗲𝗺𝗲𝗻𝘁 – 𝗡𝗲𝘁𝗦𝗰𝗶 𝗖𝗼𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝟮𝟬𝟮𝟱 📢

Joseph Lizier, Associate Professor at the School of Computer Science at the University of Sydney.

"Inferring network models from multivariate time-series data"