Package mici

MCMC samplers based on simulating Hamiltonian dynamics on a manifold.

Expand source code Browse git
# -*- coding: utf-8 -*-
""" MCMC samplers based on simulating Hamiltonian dynamics on a manifold. """

__authors__ = "Matt Graham"
__license__ = "MIT"

import mici.adapters
import mici.autodiff
import mici.integrators
import mici.matrices
import mici.samplers
import mici.solvers
import mici.stagers
import mici.states
import mici.systems
import mici.transitions

Modules

mici.adapters

Methods for adaptively setting algorithmic parameters of transitions.

mici.autodiff

Automatic differentation fallback for constructing derivative functions.

mici.autograd_wrapper

Additional autograd differential operators.

mici.errors

Exception types.

mici.integrators

Symplectic integrators for simulation of Hamiltonian dynamics.

mici.matrices

Structured matrix classes implementing basic linear algebra operations.

mici.progressbars

Progress bar classes for tracking progress of chains.

mici.samplers

Monte Carlo sampler classes for peforming inference.

mici.solvers

Solvers for non-linear systems of equations for implicit integrators.

mici.stagers

Classes for controlling sampling of Markov chains split into stages.

mici.states

Objects for recording state of a Markov chain and caching computations.

mici.systems

Hamiltonian systems encapsulating energy functions and their derivatives.

mici.transitions

Markov transition kernels.

mici.utils

Utility functions and classes.