Nice recognition by the USA Today that #Koopmanoperator based AI systems provide a framework for real time diagnostics and reasoning for network security threats! Congrats to everyone at MixMode ! #AI #ML

A big thanks to the great team!: Francisco Acosta Santi Acosta-Mendoza Nina Miolane 🦋 @ninamiolane.bsky.social Newest version of the paper: biorxiv.org/content/10.110…

Our paper "Identifying Equivalent Training Dynamics" was accepted as a spotlight (top ~5%) 💡 to NeurIPS! A real dream team (but largely Twitterless team) on this with Igor Mezic , Yannis Kevrekidis, Juan Bello-Rivas, Maria Fonoberova, Ryan Mohr 1/n arxiv.org/abs/2302.09160

Koopman Operator Theory provides a framework for #metalearning of #neuralnetworks . Our paper arxiv.org/abs/2302.09160 - accepted as #NeurIPS highlight - with Will et al. provides methodology for comparing neural network training protocols. #AI #ML #KoopmanOperator

AI for dynamical systems: the architecture that incorporates 1) a block of DNN for embeddings (feature engineering) and 2) a block of Koopman Operator based predictor/controller is emerging as a promising one in terms of accuracy and lean computation. Kolmogorov Arnold Networks

I got into science in no small part due to reading James Gleick's "#Chaos" with its wonderful depictions of the discovery process. #Lorenzattractor was one of those great discoveries. It is an example of a physical system exhibiting chaotic motion and sensitivity to initial

🚨 new paper out in Chaos "Koopman Learning with Episodic Memory"! pubs.aip.org/aip/cha/articl… Really excited about this work and, as always, a real honor to work with Igor Mezic and Aimdyn Inc. 1/10

Last May, we held a workshop in Otranto, Italy, connecting Koopman operator theory and climate science. Here is a writeup on it climateforesight.eu/articles/manif… #AI #Koopmanoperator #ML #dynamicalsystems #chaostheory

A nice application of #Koopmanoperator theory to galactic dynamics here: academic.oup.com/mnras/article/…

This paper applying Koopman to badger movement is fun, especially given my postdoc was in math biology at University of Warwick (loved it there!),biorxiv.org/content/10.110…

Foundational models collect all the data from all available instances of the process (e.g. language) to parametrize input to output relationships - interpolatively, with no real time updates. In contrast, the human brain collects data individualized to the process from relatively

Survey article: "An overview of Koopman-based control: From error bounds to closed-loop guarantees" (by Robin Strässer, Karl Worthmann, Igor Mezić, Julian Berberich, Manuel Schaller, Frank Allgöwer): arxiv.org/abs/2509.02839

The use of Koopman operator theory in control is advancing at a fast pace. Here is a survey of theoretical guarantees that establish the backbone for computations and applications arxiv.org/abs/2509.02839. #Koopmanoperator #controltheory #AI #ML

As I mentioned in another post, the advances in Koopman Operator theory and applications are happening at a fast pace, especially in AI and control applications. At the meeting, uniri.hr/en/science-and… expertly organized by Ervin Kamenar we expect a vigorous discussion that will

For all Koopmanistas, the talks from the KOT II conference are now online! Thanks to Ervin Kamenar for the organization and Marko Kozlov for his excellent work on the recordings. You can find all the talks on the YouTube Playlist: youtube.com/@KOTForum/play… #KoopmaOperator

On static Koopman operators: every ML approach, including transformers, at the final step represents output (y) as a sum of (possibly nonlinear) functions of input (x) i.e. lifting of x. In ams.org/notices/202107… I introduced the notion of linear operators associated with maps