Automatic Differentiation Toolbox
Description
Implements the forward-mode automatic differentiation for
multivariate functions using the matrix-calculus notation from Magnus and
Neudecker (2019) . Two key features of the package
are: (i) it incorporates various optimisation strategies to improve performance;
this includes applying memoisation to cut down object construction time, using
sparse matrix representation to speed up derivative calculation, and creating
specialised matrix operations to reduce computation time; (ii) it supports
differentiating random variates with respect to their parameters, targeting
Markov chain Monte Carlo (MCMC) and general simulation-based applications.