Chris Rackauckas

YouTube video by Julia Dispatch Julia Health with Jacob Zelko

#Julialang Dispatch: we welcome Jacob Zelko, who takes us on a fascinating journey from his early days as a #biomedical engineering student through his work at the CDC during the COVID-19 pandemic, and into his current exploration of applied #categorytheory.

www.youtube.com/watch?v=HyMp...

Well no one responded to libflame issues for 4 years and haven't tagged in 6. So at what point do you stop waiting and just say welp, here are the results. I thought 4 years was enough. It's on EPYC which isn't in their set I think? At least on that CPU it's not good.

BLIS is a really slow BLAS lol

Increasing spectral DCM flexibility and speed by leveraging Julia’s ModelingToolkit and automated differentiation Abstract. Using neuroimaging and electrophysiological data to infer neural parameter estimations from theoretical circuits requires solving the inverse problem. Here, we provide a new Julia language p...

This demonstrates how the optimization and Bayesian estimation of the Julia SciML tooling could be the tool analyzing your fMRIs in the near future!

See the link for the full paper:
direct.mit.edu/imag/article...

New paper on fitting #neuroscience models to neuroimaging and electrophysiological data! We show that the previous techniques within the field (Spectral DCM) can be greatly improved using automatic differentiation, #julialang #sciml, ModelingToolkit, and more.

SciML: Open Source Software for Scientific Machine Learning Open Source Software for Scientific Machine Learning

How successful has the SciML Small Grants program been at getting newcomers to contribute to open source software #oss for #sciml #julialang and #ai4science? Very!

* 13 total projects initiated
* ~90% success rate

See the blog post summary of the first year!

sciml.ai/news/2025/07...

That's why we don't do direct symbolic regression. By doing it indirect you can always regenerate data to a size more appropriate for SR convergence.

This gives a deep dive into open source sustainability models that integrates #academia while also integrating the #startup community.

In this blog post we go into detail how Dyad is designed to be synergistic with the enabling stronger open source software, giving long-term maintenance structures and show the history of how much of what has evolved in the #sciml community had been created in tandem with the goals here.

The Strategic Connection Between JuliaHub, Dyad and the Julia Open Source Community Learn how connecting the building the right software into products has actually expanded our open source community.

How does the mission of JuliaHub and the new Dyad modeling language interact with the #julialang open source (#OSS) community?

juliahub.com/blog/the-str...

Additionally, we show that this is much more capable of learning the correct equations than techniques like #SINDy as it is better able to use prior information about possible equations and model structure!

We can learn dynamical systems that capture the correct statistical properties of the original system, and use symbolic regression to sparsify the results to predicted governing equations. Use case: learning neural mass models, i.e. models of average behaviors of large brain regions

However, we can force the perturbed model to be non-chaotic, making the fitting process orders of magnitude easier. It also gives a smooth loss landscape with less or no bad local minima.

So how did we overcome this? By making a non-chaotic training of course! We instead fit a perturbed model with the curious behavior that it retains the same global optima, i.e. the correct fitting parameters of the perturbed model is also a correct fitting model of the original model.

Shadowing Methods for Forward and Adjoint Sensitivity Analysis of Chaotic Systems | FS In this post, we dig into sensitivity analysis of chaotic systems. Chaotic systems are dynamical, deterministic systems that are extremely sensitive to small changes in the initial state or the system...

Other work that we have done has highlighted that #automaticdifferentiation calculates incorrect derivatives on chaotic systems (frankschae.github.io/post/shadowi...).

How chaotic is chaos? How some AI for Science / SciML papers are overstating accuracy claims - Stochastic Lifestyle Just how chaotic are chaotic systems? Many of you may have heard of “the butterfly effect” but don’t quite know the mathematics behind such systems. What I want to demonstrate is the “sensitive depend...

What's significant here? Well in a recent blog post (www.stochasticlifestyle.com/how-chaotic-...) I described how a lot of #sciml and #ai4science did not correctly handle the long-term behavior of chaotic systems.

Our new manuscript shows how to extend automated model discovery and universal differential equations to chaotic systems in #neuroscience using a trick from control literature known as the Prediction Error Method (PEM)!

Check out the paper at: arxiv.org/abs/2507.03631

#sciml #ai4science

YouTube video by JuliaHub Getting Started with Dyad Studio - Extended Edition (Complete In-Depth Tutorial)

Check out the extended edition of our Dyad Studio tutorial! This one really dives into detail, showing you every step of using Dyad in modeling and simulation.

#julialang

www.youtube.com/watch?v=5zKL...

YouTube video by Julia Dispatch Julia Dispatch Podcast: March 2025 Newsletter Deep Dive with Chris Rackauckas & Stefan Krastanov

New Julia Dispatch Podcast! Join Chris Rackauckas and Stefan Krastanov as they explore the latest developments in the Julia programming language ecosystem from the March 2025 "This Month in Julia World" newsletter.

www.youtube.com/watch?v=i9pe...

#julialang

YouTube video by JuliaHub Getting Started with Dyad Studio

In this comprehensive tutorial, we'll walk you through everything you need to know to get started with #Dyad Studio - the next-generation modeling system that seamlessly combines GUI and textual modeling.

www.youtube.com/watch?v=QLlX...

#julialang #acausal #modelica #sciml

So if you've ever wanted to merge #acausal modeling and scientific #machinelearning, check out the Dyad documentation! help.juliahub.com/dyad/dev/

This is mixed with a complete system for doing classical controls analysis, for example auto-tuning PID controllers and performing frequency-based analyses. help.juliahub.com/dyad/dev/ana...

It is built upon the #julialang #sciml ecosystem, meaning the generated code is accessible and easily re-targetable. You can write your own backends, force the usage of #gpu acceleration, do differentiable programming, etc.

Dyad comes with #VSCode extensions to seamlessly perform static analyses for furthering model correctness by ensuring the physics is correct before code generation.

Dyad: Making Hardware as Easy as Software Dyad redefines engineering workflows with a declarative, AI-ready modeling system that bridges simulation, devops, and embedded hardware targets.

@JuliaHub_Inc
has released a new modeling language, Dyad, which integrates traditional physical modeling a la #modlica with scientific machine learning #sciml in a high-level and declarative form. Check out our release post!

juliahub.com/blog/dyad-ma...

#julialang #mbse #ai

This is the first results of the new BVP solvers to share, with many more to come soon (that will be its own publication very soon, lots of new tricks!).

For more detailed BVP solver benchmarks, see the SciMLBenchmarks: docs.sciml.ai/SciMLBenchma...

With modeling heat pumps and dehumidifiers, we were able to show that the latest boundary value problem (#BVP) solvers in #julialang #sciml outperform the Fortran wrapped bvp_solver of SciPy and the native bvp4c/5c solvers of MATLAB.

Published article here: authors.elsevier.com/c/1lHcein8Vr...

What does this do that TagBot doesn't? For the most part we seem fine with TagBot/Registrator as the release process.

YouTube video by Drumeo Mike Portnoy Hears Taylor Swift For The First Time

Artistic rendition of using #julialang on models with only 3 ODEs

www.youtube.com/watch?v=cHl_...