Adaptive probabilistic ODE solvers without adaptive memory requirements

Add to your list(s) Download to your calendar using vCal

• Nico Krämer - Danish Technical University
• Thursday 07 November 2024, 16:00-17:00
• SS03, Computer Lab.

If you have a question about this talk, please contact Carl Henrik Ek.

Adaptive probabilistic solvers for ordinary differential equations (ODEs) have made substantial progress in recent years but can still not solve memory-demanding differential equations. In this talk, I review recent developments in numerically robust fixed-point smoothers and how to use them for constructing adaptive probabilistic ODE solvers. These new algorithms use drastically less memory than their predecessors and are the first adaptive probabilistic numerical methods compatible with scientific computing in JAX .

This talk is part of the ML@CL Seminar Series series.

This talk is included in these lists:

• Hanchen DaDaDash
• ML@CL Seminar Series
• Probabilistic Numerics Reading Group
• SS03, Computer Lab
• ml@cl-math

Note that ex-directory lists are not shown.