Differentiable astrophysics at scale: solving and differentiating ODE ensembles on the GPU
astro-ph.IM, astro-ph.CO, astro-ph.GA, gr-qc, physics.comp-ph
Submitted: 2026-09-23
Updated: 2026-09-23
Code: https://github.com/ECLIPSE-AI4Science/gradsolve
Terminology
Sources
- Neural Ordinary Differential Equations
- BlackJAX: Composable Bayesian inference in JAX
- On Neural Differential Equations
- Adam: A Method for Stochastic Optimization
- torchode: A Parallel ODE Solver for PyTorch
- A Comparison of Automatic Differentiation and Continuous Sensitivity Analysis for Derivatives of Differential Equation Solutions
- MCMC using Hamiltonian dynamics
- Physics Is All You Need? A Case Study in Physicist-Supervised AI Development of Scientific Software
- Composable Effects for Flexible and Accelerated Probabilistic Programming in NumPyro
- GRADSOLVE: fast exact gradients for ODE ensembles on GPUs
- Differentiable Conservative Radially Symmetric Fluid Simulations and Stellar Winds -- jf1uids
- HIcosmo: a differentiable JAX-based framework for cosmology inference
- Differentiable N-body code for Galactic Dynamics -- Odisseo
Related papers
- A signal dedispersion algorithm for imaging-based transient searches
- AVICA: A fully automated CASA pipeline for large volume VLBI data calibration
- Spectral Map Making with SPHEREx
- Long-Integration Magnetar Burst Observatory (LIMBO): Instrument Summary and Early FRB Rate Constraints
- Towards independent event horizon imaging of the supermassive black holes in M87 and the Milky Way
- A PINK update: Improvements to the CELEBI fast radio burst data reduction and analysis pipeline