A list of principles put forth by mathematicians, for mathematicians and other researchers, regarding the use of AI in research. "Number #9 will surprise you!" leidendeclaration.ai
Who's Who of AI
Clément Canonne
Senior Lecturer #USydCompSci at the University of Sydney. Postdocs IBM Research and Stanford; PhD at Columbia. Converts ☕ into puns: sometimes theorems. He/him.
What they're sharing
Leiden Declaration on Artificial Intelligence and Mathematics leidendeclaration.ai
[2605.23225] Entropy Equivalence Testing arxiv.org
Near-Optimal Pure Machine Unlearning for Smooth Strongly Convex Losses arxiv.org
View on Bluesky →
Optimal Rates for Differentially Private Hypothesis Testing with E-values arxiv.org
View on Bluesky →
arxiv.org
blog.geomblog.org
AI is a Meteor. Don’t be a Dinosaur. windowsontheory.org
cacm.acm.org
View on Bluesky →
Articles & links
New preprint up, led by Joy (Qiping) Yang and Yash Pote, with Jonathan Scarlett: we consider a relaxation of the standard (distribution) closeness testing task, where the algo now only needs to distinguish between p=q (equal) and |H(p)-H(q)|=Ω(1) (their entropies differ). 📝 ar…
Clément Canonne reposted
Near-Optimal Pure Machine Unlearning for Smooth Strongly Convex Losses Matthew Regehr, Gautam Kamath, Andrew Lowy http://arxiv.org/abs/2606.01527
Clément Canonne reposted
Optimal Rates for Differentially Private Hypothesis Testing with E-values Ben Jacobsen, Tomas Gonzales, Gavin Brown, Kassem Fawaz, Aaditya Ramdas http://arxiv.org/abs/2605.28952
New preprint online: on multiple trade-offs b/w communication, sample, and randomness complexities for the fundamental task of Gaussian mean testing in a distributed setting. Led by the brilliant Nimitt (undergrad at IIT Gandhinagar) during a remote internship last year! 📝 arx…
Good blog post by Suresh Venkatasubramanian (@geomblog.bsky.social) on the latest installment of the Spectacle regarding Anthropic and the US government. He brings the popcorn, but the receipts and analysis, too. blog.geomblog.org/2026/06/the-...
Boaz is incredibly smart! Many are! Do they not realize that "I'm happily building a thing I genuinely believe can be compared to a planetary-scale catastrophe,* I suggest you make peace with it" is NOT great PR? (*Regardless of whether you/I believe it) windowsontheory.org/20…
Clément Canonne reposted
1961: CACM publishes Tony Hoare's quicksort algorithm. 2026: CACM publishes a self-promo article by some random guy who's pitting different LLMs against each other to pump unmitigated slop into a GitHub repo. Is this what my membership money's going toward, @acm.org?
Their own posts
Recent commentary
The OpenAI breakthrough on the unit distance problem seems genuinely impressive to my semi-layman eye. (It is!) One possible takeaway, though, is that if you throw an incredible amount of money and resources at focused research, there WILL be impressive progress. I wish we tried that, too.
I uploaded a picture of a random root vegetable and asked ChatGPT to identify it. It told me it was a stochastic carrot.
"AI will replace U, they said" "Nothing yoai can do aboait it. It's the faitaire."
The Pope taking such a strong stance on AI feels rather unorthodox
I don't know what's happening with Google Scholar,* but lately I have been receiving a large number of absurd Google Scholar Alerts for papers anything from one to 15 years old. Is that widespread? *My fear is they offloaded GScholar to AI agents, and it's on a steady way to becoming unusable.
Their network
In Clément Canonne's orbit
Center = Clément Canonne. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.