Skip to content

A library for differentiating tacit BQN expressions.

Notifications You must be signed in to change notification settings

jacobusmmsmit/tacitdiff

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 

Repository files navigation

TacitDiff

A library for differentiating tacit BQN expressions in different ways.

Currently supported:

  • Symbolic differentiation (by saltysylvi)
  • Tape-based reverse-mode algorithmic differentiation

Planned features:

  • Forward-mode algorithmic differentiation
  • (Accurate) finite differencing

Why do this?

I wanted to learn more about what goes into writing and optimising an autodiff library. I also wanted to challenge myself to come up with a recursive algorithm for generating and evaluating Wengert lists aka gradient tapes. BQN is a very designed language and autodiff fits well into its array programming paradigm. Finally, I just wanted to write something bigger than a few one liners in BQN.

About

A library for differentiating tacit BQN expressions.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published