jax.lax.linalg.cholesky_update#

jax.lax.linalg.cholesky_update(r_matrix, w_vector)[source]#

Cholesky rank-1 update.

Given a Cholesky decomposition \(A = R.T \, R\) and a vector \(w\), computes the Cholesky decomposition of \(A + w \, w.T\) in \(O(N^2)\) time.

Parameters:
  • r_matrix (ArrayLike) – An upper-triangular matrix (R) such that \(A = R^T \, R\).

  • w_vector (ArrayLike) – A vector \(w\) for rank-1 update.

Returns:

A new upper-triangular matrix \(R\) defining the Cholesky decomposition of \(A + w \, w^T\).

Return type:

Array