RustNL

Manuel Drehwald -- Computing derivatives (gradients, jacobians, hessians, …) is relevant for fields like Machine Learning or scientific computing, where we have function parameters that we want to optimize. “Automatic Differentiation” allows us to apply this concept from calculus to real-world code, written in languages like Rust. We are currently working hard on enabling the new std::autodiff macro for nightly. This macro is based on a very performant, but experimental LLVM project called Enzym

Player unavailable or embedding blocked? Use the YouTube link above.