Skip to content

Latest commit

 

History

History
 
 

Ops for numerical methods.

This module provides ops to perform various mathematical tasks commonly needed when building quantitative finance models. We do not aim to provide exhaustive coverage here. Tensorflow and Tensorflow Probability provide a significant suite of methods already and the methods here are meant to build on those.

Some of the modules/functions provided are:

  • math.interpolation: Ops to perform linear and cubic interpolation.
  • math.optimizer: Ops for numerical optimization.
  • math.pde: Ops to numerically solve partial differential equations using finite difference methods. Currently, only linear second order PDEs are supported as this is the most commonly needed case.
  • math.random: Ops to compute low discrepancy sequences.
  • math.root_search: Provides the Brent method for computing roots of functions in one dimension.
  • math.segment_ops: Utility methods to apply some element wise ops in a segment.