- Evaluating Derivatives: Principles and Techniques of Algorithmic Differentiation, Second Edition: clear exposition of the thoery and practice of forward and reverse mode automatic diffentation
- David Gay's RAD library: see dep/rad folder into this repository,
- Adept: A fast automatic differentiation library for C++: one of the fastes implementations of reverse mode AD in C++ via operator overloading and template metaprogramming
- Stan: framework for frequentist and bayesian inference of statistical distribution-based models, includes reverse mode automatic differentiation engine (not explicitely documented)
- Understand reverse mode logic, convert/extend RAD library to Lua
- Implement efficient Lua version of reverse mode differentiation based on Adept and Stan
- Optimize implementation of point 2 for LuaJIT/FFI