priors.py
# Copyright 2016 Valentine Svensson, James Hensman, alexggmatthews
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from __future__ import absolute_import
import tensorflow as tf
import numpy as np
from .param import Parameterized
from . import densities
from ._settings import settings
float_type = settings.dtypes.float_type
np_float_type = np.float32 if float_type is tf.float32 else np.float64
class Prior(Parameterized):
def logp(self, x):
"""
The log density of the prior as x
All priors (for the moment) are univariate, so if x is a vector or an
array, this is the sum of the log densities.
"""
raise NotImplementedError
def sample(self, shape=(1,)):
"""
a sample utility function for the prior.
"""
raise NotImplementedError
def __str__(self):
"""
A short string to describe the prior at print time
"""
raise NotImplementedError
class Gaussian(Prior):
def __init__(self, mu, var):
Prior.__init__(self)
self.mu = np.atleast_1d(np.array(mu, np_float_type))
self.var = np.atleast_1d(np.array(var, np_float_type))
def logp(self, x):
return tf.reduce_sum(densities.gaussian(x, self.mu, self.var))
def sample(self, shape=(1,)):
return self.mu + np.sqrt(self.var)*np.random.randn(*shape)
def __str__(self):
return "N("+str(self.mu) + "," + str(self.var) + ")"
class LogNormal(Prior):
def __init__(self, mu, var):
Prior.__init__(self)
self.mu = np.atleast_1d(np.array(mu, np_float_type))
self.var = np.atleast_1d(np.array(var, np_float_type))
def logp(self, x):
return tf.reduce_sum(densities.lognormal(x, self.mu, self.var))
def sample(self, shape=(1,)):
return np.exp(self.mu +
np.sqrt(self.var)*np.random.randn(*shape))
def __str__(self):
return "logN("+str(self.mu) + "," + str(self.var) + ")"
class Gamma(Prior):
def __init__(self, shape, scale):
Prior.__init__(self)
self.shape = np.atleast_1d(np.array(shape, np_float_type))
self.scale = np.atleast_1d(np.array(scale, np_float_type))
def logp(self, x):
return tf.reduce_sum(densities.gamma(self.shape, self.scale, x))
def sample(self, shape=(1,)):
return np.random.gamma(self.shape, self.scale, size=shape)
def __str__(self):
return "Ga("+str(self.shape) + "," + str(self.scale) + ")"
class Laplace(Prior):
def __init__(self, mu, sigma):
Prior.__init__(self)
self.mu = np.atleast_1d(np.array(mu, np_float_type))
self.sigma = np.atleast_1d(np.array(sigma, np_float_type))
def logp(self, x):
return tf.reduce_sum(densities.laplace(self.mu, self.sigma, x))
def sample(self, shape=(1,)):
return np.random.laplace(self.mu, self.sigma, size=shape)
def __str__(self):
return "Lap.("+str(self.mu) + "," + str(self.sigma) + ")"
class Beta(Prior):
def __init__(self, a, b):
Prior.__init__(self)
self.a = np.atleast_1d(np.array(a, np_float_type))
self.b = np.atleast_1d(np.array(b, np_float_type))
def logp(self, x):
return tf.reduce_sum(densities.beta(self.a, self.b, x))
def sample(self, shape=(1,)):
return np.random.beta(self.a, self.b, size=shape)
def __str__(self):
return "Beta(" + str(self.a) + "," + str(self.b) + ")"
class Uniform(Prior):
def __init__(self, lower=0, upper=1):
Prior.__init__(self)
self.log_height = - np.log(upper - lower)
self.lower, self.upper = lower, upper
def logp(self, x):
return self.log_height * tf.cast(tf.size(x), float_type)
def sample(self, shape=(1,)):
return (self.lower +
(self.upper - self.lower)*np.random.rand(*shape))
def __str__(self):
return "U("+str(self.lower) + "," + str(self.upper) + ")"