JAX scientific ML ecosystem:
Probably the reason you're here. I would highlight:
-
Equinox: elegant neural networks.
-
Diffrax: numerical ODE/SDE solvers.
-
jaxtyping: shape/dtype annotations for arrays.
-
Lineax: linear/least-squares solvers. (new!)
-
Optimistix: root finding, least squares, etc. (new!)
Other links:
- Twitter:
- Google scholar: here
- Personal website: kidger.site
- Neural ODE/SDE textbook: arXiv/2202.02435
Me:
I currently wear multiple hats across bio/ML/CS at Cradle.bio. These days I am generally interested in scientific ML, and specifically the application of ML to unsolved problems in biology!
I also hold an honorary lectureship at Imperial College London. In past lives I previously wore the same multitude of hats at Google X, and did my PhD at the University of Oxford.