diabayes.SVI.compute_phi#

diabayes.SVI.compute_phi(x: Float[Array, 'N M'], gradp: Float[Array, 'N M'], gradq: Float[Array, 'N M']) → Float[Array, 'N M'][source]#

Compute the Stein variational gradients for a set of particles (x) and the gradients of the log-likelihood function (gradp) and the log-prior (gradq).

Parameters:

  • x (ParticleArray) – The set of invertible parameters (“particles”) of shape (Nparticles, Ndimensions)

  • gradp (GradientArray) – The gradients of the log-likelihood (gradp) and the log-prior (gradq) with respect to the invertible parameters. Has a shape (Nparticles, Ndimensions)

  • gradq (GradientArray) – The gradients of the log-likelihood (gradp) and the log-prior (gradq) with respect to the invertible parameters. Has a shape (Nparticles, Ndimensions)

Returns:

grad_x – The directional gradients for the particle updates, e.g. new_x = x + step_size * grad_x. Has the same shape as x and gradp, gradq.

Return type:

Gradients

On this page