Self-Scaled Broyden Family of Quasi-Newton Methods in JAX
This is an incremental technical note documenting an implementation for JAX users.
The authors implemented the Self-Scaled Broyden family of quasi-Newton methods in JAX, including variants like BFGS and DFP, to facilitate adoption within the JAX community.
We present a JAX implementation of the Self-Scaled Broyden family of quasi-Newton methods, fully compatible with JAX and building on the Optimistix~\cite{rader_optimistix_2024} optimisation library. The implementation includes BFGS, DFP, Broyden and their Self-Scaled variants(SSBFGS, SSDFP, SSBroyden), together with a Zoom line search satisfying the strong Wolfe conditions. This is a short technical note, not a research paper, as it does not claim any novel contribution; its purpose is to document the implementation and ease the adoption of these optimisers within the JAX community. The code is available at https://github.com/IvanBioli/ssbroyden_optimistix.git.