pypomp.functional.mop

pypomp.functional.mop(struct: PompStruct, thetas_array: Array, J: int, alpha: float, keys: Array) → Array[source]

This is a pure functional implementation of the MOP differentiable particle filter, intended for users who need to compose it within custom JAX loops or higher-order functions. For a more user-friendly (but impurely-functional) interface to train models using the MOP objective, see pypomp.core.pomp.Pomp.train().

Unlike the standard particle filter (pypomp.functional.pfilter()), the MOP objective is specifically designed to be fully differentiable with respect to the model parameters. This allows for the computation of gradients and Hessians of the log-likelihood using JAX’s automatic differentiation capabilities.

This function evaluates the log-likelihood for the given parameter sets, but it is primarily intended to be used as an objective function within gradient-based optimization routines (e.g., pypomp.functional.train()).

Parameters:

  • struct (PompStruct) – The compiled structural representation of the POMP model.

  • thetas_array (jax.Array) – Array of initial parameters. Shape (n_reps, n_params). Must be aligned with the canonical order of struct.param_names (e.g. prepared via align_params).

  • J (int) – Number of particles.

  • alpha (float) – Alpha parameter for MOP.

  • keys (jax.Array) – Random keys. Shape (n_reps, …).

Returns:

Negative MOP log-likelihood estimates.

Return type:

jax.Array

Note

To align and stack input parameter dictionaries/scalars into the correct canonical ordering required by these arrays, you can use pypomp.functional.align_params().