New blog post: Some Comments on “Magnifica Humanitas” In which I comment on the recent encyclical by Pope Leo XIV from the perspective of a machine learning researcher. bastian.rieck.me/blog/2026/ma...

Is it an unwritten rule that frontier AI companies have to pick their name out of Tolkien's legendarium? What comes after "peak Tolkien?" Star Wars? Game of Thrones? Something else?

The hike vs helicopter framing assumes choice. For dyslexics & ADHDers like me, AI is the wheelchair, not the helicopter; not skipping the climb, making the climb possible. Your 'cannot abdicate responsibility' line holds. carlmfreeman.substack.com/p/nobody-kno...

New blog post: Scholarship in the Age of AI In which I use the recent debate on arXiv policies to ponder what it means to be a scholar now. Feedback always welcome! #Academia #AI bastian.rieck.me/blog/2026/sc...

After a certain age, discovering new music gets harder. Siri didn't understand my query, said "Hold my beer!", and served me "Birdie" by "NOVA Rei." (Can you guess what I wanted to listen to originally?)

Did anyone else notice a move of generic AI slop to platforms like Zenodo? I used to get endorsement requests for arXiv, which I did not feel comfortable with, and now I am noticing more and more "weird stuff" (consciousness + topology + quantum whatever) hosted on Zenodo.

Gladiator vibes...good evening, y'all!

Some respite before the heat starts again. Good morning, everyone!

Remember when I was first becoming a Jesuit and gave away a bunch of my stuff? Well here's your chance again! All of this is free to a good home if you pay shipping. (Note the Majora's Mask has a cracked case.) DM me if you are interested. Serious inquiries only! Needs to go in the next week-ish.

📣 LOGML'26 Speaker Series 🎤 Soledad Villar @soledadvillar.bsky.social (Johns Hopkins) Her research spans equivariant ML, GNNs, with applications to computational biology. 📍 Imperial College London 📅 13–17 July 2026 🔗 www.logml.ai #LOGML #GraphML #GeometricDeepLearning #SummerSchool

The Boston TAG Party (August 18–20) is open for submissions! openreview.net/group?id=TAG... Submit your work in topology, algebra, and geometry.

Every ad now

Papers are tactics. Grants are strategy. — Sun Tzu, thinking about CS conference outcomes.

Walking to the office always gives me Lord of the Rings vibes. #Fribourg #Freiburg

If you want to know a bit more about my recent paper with Juan, Ernst and Bastian, watch this video of Juan explaining it :)

A huge LEAP forward for graph learning and topological deep learning: Juan is taking about our recent ICLR work on learnable positional encodings. #MachineLearning #ICLR2026 🎥: youtu.be/7aniG7esPWU... 📝: openreview.net/forum?id=8XF...

The fake expert voice of LLMs may slowly erode calibrated thinking. New blog on epistemic humility and why there are no shortcuts. bastian.rieck.me/blog/2026/hu...

A special treat for the students in my course „Introduction to Topological Data Analysis“: Bastian Rieck (@pseudomanifold.topology.rocks) gave a very interesting guest lecture on topological methods in deep learning.

@annajobin.bsky.social"Auch die Bereitstellung von Infrastruktur ist eine klassische Aufgabe des Service public, die private Medien nicht leisten müssen. Die SRG könnte also stärker die Funktion wahrnehmen, neue Technologien zu testen und vorbildliche Standards zu setzen." www.nzz.ch/schweiz/die-...

Multiple positions (open-rank) for a Professor in Data-Driven Materials Discovery are open @unifr.bsky.social. jobs.sciencecareers.org/job/677463/p... Join us here, it's great! (Please retweet for additional reach!) #Academia

Interested in the analysis of GNN expressive power, limitations of the WL test or more in general in the definition of complete metrics on graphs? 👀 Check out this super cool work led by @martin.topology.rocks !🔥

We find it does great on non-isomorphic graph distinction by evaluating on BREC icml.cc/virtual/2024... and as an inductive bias on ZINC (compared to other homomorphism and subgraph count methods)

Not all is lost because we can approximate our true metric by sampling computable graphs (of a nice treewidth) and also homomorphisms!

Under mild assumptions, we prove our metric: 1.  Is 𝒸𝑜𝓂𝓅𝓁𝑒𝓉𝑒 regarding the isomorphic distinction. 2.  Is 𝓅𝑒𝓇𝓂𝓊𝓉𝒶𝓉𝒾𝑜𝓃 𝒾𝓃𝓋𝒶𝓇𝒾𝒶𝓃𝓉. 3. Can 𝒹𝒾𝓈𝓉𝒾𝓃𝑔𝓊𝒾𝓈𝒽 induced subgraphs (even if a retraction exists)! 4. Has a 𝒻𝒾𝓃𝑒𝓇 𝓉𝑜𝓅𝑜𝓁𝑜𝑔𝓎 than the metric space induced by WL. However... It's hard to compute :(

Inspired by the 𝘎𝘳𝘰𝘮𝘰𝘷-𝘏𝘢𝘶𝘴𝘴𝘥𝘰𝘳𝘧 distance, we define distortion as the maximum difference of the attributes of a graph 𝗚 and a graph 𝗚' through a homomorphism. We construct a metric from this 𝘥𝘪𝘴𝘵𝘰𝘳𝘵𝘪𝘰𝘯 by taking the homomorphism providing the infimum result and looking at both directions!

To kick off the PhD journey with @pseudomanifold.topology.rocks: What are the limitations of the WL metric, and what is an 𝘪𝘯𝘧𝘰𝘳𝘮𝘢𝘵𝘪𝘷𝘦 𝘮𝘦𝘵𝘳𝘪𝘤? We answer these questions with our 𝗚𝗿𝗮𝗽𝗵 𝗛𝗼𝗺𝗼𝗺𝗼𝗿𝗽𝗵𝗶𝘀𝗺 𝗗𝗶𝘀𝘁𝗼𝗿𝘁𝗶𝗼𝗻 arxiv.org/abs/2511.03068 @olgatticus.bsky.social, Kavir and @erikjbekkers.bsky.social

Two new videos on #topology 🍩 and #DeepLearning 🤖 just dropped: - Part 1: youtu.be/p1GKAV0DAPI - Part 2: youtu.be/_b19lhGRJmg Both are part of my invited talk at the "OneMath World School 2026." Thanks for having me!

The Conference of Catholic Bishops is... not playing around in this brief. www.supremecourt.gov/DocketPDF/25...

Nice and quite nuanced. And I love this quote!

New blog post: The Better Lesson? Geometry and Topology in the Era of Deep Learning In which I think a bit about the role fundamental mathematical concepts might yet have to play in the quest for better deep learning architectures. Comments welcome! bastian.rieck.me/blog/2026/be...

Have some (applied) topology research you want to share with the world? Consider adding it to 🍩DONUT🍩, our Database of Original & Non-Theoretical Uses of Topology. Great intro by Barbara Giunti: www.youtube.com/watch?v=eQ_5... 🏠 donut.topology.rocks

Very proud of my student Juan Amboage, who managed to get his Master thesis accepted at #ICLR2026, with significant help by Ernst Röell and @pseudomanifold.topology.rocks , and very very little input from my side. If you want to know more, here is the ArXiv version: arxiv.org/abs/2510.00757

📢 The submission deadline for ✨GRaM workshop✨ at @iclr-conf.bsky.social is 5th February 2026. 🌟Submit your exciting works in Geometry-grounded representations. We welcome submissions in multiple tracks i.e. 📄 Proceedings 📝extended abstract 👩‍🏫Tutorial/blogpost as well as an exciting challenge!

AI is a new idea? Not quite! In this short blog post, geared towards a general audience, I take a brief tour through history, starting with Homer and ending with ChatGPT. Enjoy! bastian.rieck.me/blog/2025/new/

For more details check out our website: http:// gram-workshop.org or follow @gram-org.bsky.social

GRaM 2.0 is being organised by an ✨incredible team✨ @erikjbekkers.bsky.social @alisometry.bsky.social @sharvaree.bsky.social @oumarkaba.bsky.social @jmtomczak.bsky.social @sukjulian.bsky.social Hannah Lawrence, Robin Walters, Manuel Lecha, Elizabeth Baker and Stefanie Jegelka.

📢The second edition of ✨GRaM workshop✨ is here this time at #ICLR26. 🌟Submit your exciting works in Geometry-grounded representations. We welcome submissions in multiple tracks i.e. 📄 Proceedings 📝extended abstract 👩‍🏫Tutorial/blogpost as well as an exciting challenge!

Geometry & topology may not save the world, but they may well save something that is homotopy-equivalent to it.

Celebrating the author contributions of our ELLIS Members and PhD Students at @neuripsconf.bsky.social 2025 - congratulations to all authors! NeurIPS is joined for the first time by @euripsconf.bsky.social in CPH 🇩🇰, a NeurIPS-endorsed European venue co-developed by ELLIS. 🔗 https://bit.ly/48cRAib

My thesis project embarks on this question, under the supervision of both @erikjbekkers.bsky.social , @pseudomanifold.topology.rocks and @olgatticus.bsky.social we devise a metric over graphs, 𝘵𝘩𝘳𝘰𝘶𝘨𝘩 the maps that exists between them. It also has a cool name! 𝐇𝐨𝐦𝐨𝐦𝐨𝐫𝐩𝐡𝐢𝐬𝐦 𝐃𝐢𝐬𝐭𝐨𝐫𝐭𝐢𝐨𝐧 (ⒽⒹ)

All these insights resulted in a 9.5/10 thesis grade (whatever that means), which I hope serves as a hook to give it a read! 📖 On arxiv: arxiv.org/abs/2511.03068

Under mild assumptions, we obtain a metric and we show 𝘤𝘰𝘮𝘱𝘭𝘦𝘵𝘯𝘦𝘴𝘴. However, as all graphs enthusiasts might notice, this implies the GI problem. That's not the end of the story because we show that we can sample graphs and achieve completeness in expectation :0 !

Two month flew by since the start of the PhD journey with @pseudomanifold.topology.rocks and now the time is right to share how our collaborations officially started, with a question: How can we measure difference in (attributed) graphs beyond binary isomorphism testing ? 🌀

Beware of low-hanging fruits in research, lest you eat the forbidden fruit and get expelled from paradise. #showerthoughts #research #academia

Excited to be at #SfN25! Come catch my poster “Neural dimensionality expands over the course of brain-computer interface learning” wed PM (W15) if you’re interested in non-invasive BCIs, neural manifolds, human learning, etc. Building upon our recent work (doi.org/10.1101/2025.03.29.646109)

Big graph, little compute? In our #NeurIPS2025 paper we propose geometry-aware edge-contraction-based pooling methods for GNNs. Our methods preserve graph structure, make interpretable pooling choices, and ensure robust performance at downstream tasks. 🧵1/n

Now with an excellent YouTube video by @ben300694.bsky.social: youtu.be/Vkl6dLKnnJc

New blog post: Skills for the Future A brief reflection on what skills may (hopefully) remain relevant in the new (hopefully bright) future of LLMs et al. #AcademicWriting #AcademicChatter bastian.rieck.me/blog/2025/sk...