HeyJaredMachine Learning & Simulation
Explaining topics of 🤖 Machine Learning & 🌊 Simulation with intuition, visualization and code. ------ Hey, welcome to my channel of explanatory videos for Machine Learning & Simulation.
- Indexed videos, last 90 days
- 5
- Latest publication
- Aug 26, 2026
- Audience
- ~34.4K subscribers
- Earliest in this view
- Aug 4, 2026
Latest videos
My PhD Oral Defense Presentation: From Numerical Simulators to Neural Emulators and Back (opens the original)
Read excerpt
On July 29 2026, I graduated with a PhD in computer science, as presented here. In it, I synthesize my three major publications (APEBench, PRDP, Emulator Superiority) into a single story about the different roles of numerical solvers in the holistic emulation pipeline. You can read the thesis here: https://arxiv.org/abs/2608.24547 Slides are here: https://fkoehler.site/files/phd_defense_slides.pdf --- Timestamps: 00:00 Intro & why simulation is important 01:05 How simulation is done classically
Neural-Hybrid Correctors with Solver-in-the-loop in JAX (opens the original)
Read excerpt
Part 2 of the hands-on session, and where the lecture and the previous two notebooks come together: a simplified reproduction of the ML-accelerated CFD idea, built in JAX. You can find all the material here: https://github.com/Ceyron/hybridization-in-jax Recorded for the workshop "Machine Learning and Automatic Differentiation in JAX for Scientific Computing", University of Strasbourg, June 2026. ── The series ── Lecture — The Hybridization of Solvers and Deep Learning: https://youtu.be/carwzAOf
Data Assimilation and Inverse Problems with Differentiable Solvers in JAX (opens the original)
Read excerpt
Part 1 of the hands-on session: using the differentiability of Exponax to solve two inverse problems on the 2D Kolmogorov flow. You can find all the material here: https://github.com/Ceyron/hybridization-in-jax Recorded for the workshop "Machine Learning and Automatic Differentiation in JAX for Scientific Computing", University of Strasbourg, June 2026. ── The series ── Lecture: The Hybridization of Solvers and Deep Learning: https://youtu.be/carwzAOfuPE Part 0: Exponax and JAX for Kolmogorov fl
Introduction to Exponax and JAX for Kolmogorov Flow (opens the original)
Read excerpt
Part 0 of the hands-on session: getting comfortable with Exponax, a Fourier pseudo-spectral PDE solver suite written in JAX. We set up a 2D Kolmogorov flow in stream-function–vorticity form, warm it up onto the turbulent manifold, and roll it out. You can find all the material here: https://github.com/Ceyron/hybridization-in-jax Recorded for the workshop "Machine Learning and Automatic Differentiation in JAX for Scientific Computing", University of Strasbourg, June 2026. ── The series ── Lecture
Hybridization of Neural Networks and Numerical Solvers in JAX with Differentiable Physics (opens the original)
Read excerpt
This is the lecture I gave at the ML & AD for Scientific Computing in JAX workshop at Strasbourg in June 2026. All the material can be found here: https://github.com/Ceyron/hybridization-in-jax What the talk covers: - Numerical simulators and neural networks side by side: compute graphs, autodiff granularity (scalar / BLAS / algebra / whole-operation), and what actually differs - A deliberately strict definition of hybridization, and why PINNs, Neural ODEs, Hamiltonian/Lagrangian NNs and Deep Eq
Publishing over time
Last 90 days. Choose a month to open its work.
Recurring subjects
Named in the text we hold. One piece can cover several.
Audience
~34.4K subscribers
Measured Sep 19, 2026
Source's subscribers, not the number who saw an individual piece.
About this data
Counts cover the work we have indexed. Tone needs enough text and a confident classification. Excerpts and episode notes are not full articles or transcripts.
Identity or attribution wrong? Suggest a correction.