TransportMaps.Diagnostics.Routines

Module Contents

Functions

compute_vals_variance_approx_kl(d1, d2[, params1, ...])

Compute values necessary for the evaluation of the variance diagnostic variance_approx_kl()

variance_approx_kl(d1, d2[, params1, params2, ...])

Variance diagnositc
TransportMaps.Diagnostics.Routines.compute_vals_variance_approx_kl(d1, d2, params1=None, params2=None, x=None, mpi_pool_tuple=(None, None), import_set=set())[source]

Compute values necessary for the evaluation of the variance diagnostic variance_approx_kl()

Returns:

(tuple [2] ndarray [\(m\)]) –

computed values of \(\log\pi_1\) and \(\log\pi_2\)

See also

variance_approx_kl()

TransportMaps.Diagnostics.Routines.variance_approx_kl(d1, d2, params1=None, params2=None, vals_d1=None, vals_d2=None, qtype=None, qparams=None, x=None, w=None, mpi_pool_tuple=(None, None), import_set=set())[source]

Variance diagnositc

Statistical analysis of the variance diagnostic

\[\mathcal{D}_{KL}(\pi_1 \Vert \pi_2) \approx \frac{1}{2} \mathbb{V}_{\pi_1} \left( \log \frac{\pi_1}{\pi_2}\right)\]
Parameters:

  • d1 (Distribution) – distribution \(\pi_1\)

  • d2 (Distribution) – distribution \(\pi_2\)

  • params1 (dict) – parameters for distribution \(\pi_1\)

  • params2 (dict) – parameters for distribution \(\pi_2\)

  • vals_d1 (ndarray [\(m\)]) – computed values of \(\log\pi_1\)

  • vals_d2 (ndarray [\(m\)]) – computed values of:math:logpi_2

  • qtype (int) – quadrature type to be used for the approximation of \(\mathbb{E}_{\pi_1}\)

  • qparams (object) – parameters necessary for the construction of the quadrature

  • x (ndarray [\(m,d\)]) – quadrature points used for the approximation of \(\mathbb{E}_{\pi_1}\)

  • w (ndarray [\(m\)]) – quadrature weights used for the approximation of \(\mathbb{E}_{\pi_1}\)

  • mpi_pool_tuple (tuple [2] of mpi_map.MPI_Pool) – pool of processes to be used for the evaluation of d1 and d2

  • import_set (set) – list of couples (module_name,as_field) to be imported as import module_name as as_field (for MPI purposes)

Note

The parameters (qtype,qparams) and (x,w) are mutually exclusive, but one pair of them is necessary.