Hello! I'm also working on automatic differentiation crate

[djmaxus]

Hi, my name is Max. I do academic research on Reduced Order Modeling of dynamical systems and somewhat related developments for a company.

I could not find another ways to contact you, so this will be my 'email' 😃

This summer I started learning Rust and making automatic differentiation as my first project in that language.

So it was barely possible to miss your crate autodiff, which I was glad to have a look at.

For me, it's a pleasure to meet people with your background and expertise by accident.

My crate is called autodj. I would be glad for your feedback on it in case that's possible.

Best wishes,

Maksim Elizarev djmaxus.github.io

[elrnv]

Hi Max!

Thanks for your "email" :) You can often find emails of authors in Cargo.toml files in Rust projects, and in git commit history by running git log on a project.

I like your crate, and thank you for your contribution to scientific computing in Rust! There are a number of other automatic differentiation libraries (e.g. hyperdual and Fwd:AD) that have implemented more feature-full alternatives to autodiff, especially when it comes to multivariate automatic differentiation.

In this space I would personally love to see more benchmarks and efficient implementations of reverse-mode automatic differentiation. I think these can help guide future development in this area.

Best, Egor