Appears to be a breakthrough result but no mention of Nagarajan (Vaishnavh Nagarajan ) and his insight into pitfalls of next token prediction. arxiv.org/abs/2508.08222

Mikhail Gromov, the giant of geometry. First time we met when I accidentally ran into him in the Parisian underground sixteen years ago. He somehow knew about my existence and agreed to meet the following day, dedicating to me 3 hours. Today, he came to my talk in Cambridge.

Single-cell data can reveal hierarchical patterns in organismic development but popular embedding approaches often distort them. We introduce Contrastive Poincaré Maps, a self-supervised hyperbolic encoder that preserves hierarchies, scales efficiently, and uncovers developmental

"Playing with Shape and Form gives a playful and intuitive introduction to topology, favouring simple and crisp illustrations over long explanations..." - Amy Boyle Links in comments.

Day 2 of the Southeastern Sectional Meeting was full of special sessions, invited speakers, raffles, and more networking! If you attended, what's been a highlight for you?

Algebraic topology helps computers understand the "shape" of complex data. In machine learning, this is called Topological Data Analysis (TDA). TDA finds clusters, loops, and holes in datasets, revealing hidden patterns that traditional methods miss. This is used to create more

Very nice talk by Boris Hanin at the American Mathematical Society SE Sectional Meeting!

Found a better version (thanks to a follower of mine) of the primer (this time, the complete 281 pages version) on categorical logic by the great Steve Awodey. Logicians, mathematicians and computer scientists will find this one very interesting. github 🔗 in the comments

Category theory is a mathematical language for describing how systems are built by composing parts. In machine learning, it offers a powerful framework for understanding the architecture of neural networks, viewing layers as functions that compose together. It also inspires

when it rains, it pours! for years, it seemed like the ML community had lost interest in PAC learning automata and formal languages the topic had seemed "exhausted" -- mainly because essentially any reasonable thing you'd want to do was proven to be computationally hard in some

New Followers: A tutorial with no prerequisite in differential geometry required: "An Elementary Introduction to Information Geometry" 👉 tinyurl.com/ElementaryIG/

In "On Growth and Form," published in 1917, the scientist D’Arcy Thompson highlighted similarities between living and nonliving matter. His thesis — that physical and mechanical forces shape organisms — is coming back into vogue. quantamagazine.org/genes-have-har…

Beautiful piece in the AMS notices

instead of trying to get AI to prove the Big Conjecture, here's something you can do right now with high odds of success... the hidden 💎 search for math papers by profs at a top uni that are >5 years out and have <10 citations. pull up a stack of 'em. feed the stack to

Singular learning theory is a theory of machine learning of singular models, either non-identifiable models or having degenerate Fisher information matrix. Regular non-singular statistical models are handled as manifolds and singular models as algebraic varieties

[LG] The Geometry of Reasoning: Flowing Logics in Representation Space Y Zhou, Y Wang, X Yin, S Zhou... [Duke University] (2025) arxiv.org/abs/2510.09782

I sent an email to P. Diaconis yesterday to tell him how much I liked his paper ‘Generating a Random Permutation with Random Transpositions’ (1981), and that our work below was inspired by it. He actually replied, which is awesome! He also pointed me to another application of

"Mathematical Structures: From Linear Algebra over Rings to Geometry with Sheaves" link.springer.com/book/10.1007/9…

One sure way to scare people this Halloween

when i presented clrs at stanford, a fun question arose that i didn't get anywhere else: 'would pre-training a gnn to execute algorithms help on an ogb task?' it... took a while before this got checked, but thanks to Jason (my cambridge student), we now know the answer is yes 😄