tfp.distributions.TransformedDistribution

See Stable See Nightly

View source on GitHub

A Transformed Distribution.

Inherits From: Distribution

tfp.distributions.TransformedDistribution(
    distribution, bijector, kwargs_split_fn=_default_kwargs_split_fn,
    validate_args=False, parameters=None, name=None
)

Used in the notebooks

A TransformedDistribution models p(y) given a base distribution p(x), and a deterministic, invertible, differentiable transform, Y = g(X). The transform is typically an instance of the Bijector class and the base distribution is typically an instance of the Distribution class.

A Bijector is expected to implement the following functions:

• forward,
• inverse,
• inverse_log_det_jacobian.

The semantics of these functions are outlined in the Bijector documentation.

We now describe how a TransformedDistribution alters the input/outputs of a Distribution associated with a random variable (rv) X.

Write cdf(Y=y) for an absolutely continuous cumulative distribution function of random variable Y; write the probability density function pdf(Y=y) := d^k / (dy_1,...,dy_k) cdf(Y=y) for its derivative wrt to Y evaluated at y. Assume that Y = g(X) where g is a deterministic diffeomorphism, i.e., a non-random, continuous, differentiable, and invertible function. Write the inverse of g as X = g^{-1}(Y) and (J o g)(x) for the Jacobian of g evaluated at x.

A TransformedDistribution implements the following operations:

• sample Mathematically: Y = g(X) Programmatically: bijector.forward(distribution.sample(...))

• log_prob Mathematically: `(log o pdf)(Y=y) = (log o pdf o g^{-1})(y)

                 + (log o abs o det o J o g^{-1})(y)`
  

  Programmatically: (distribution.log_prob(bijector.inverse(y)) + bijector.inverse_log_det_jacobian(y))

• log_cdf Mathematically: (log o cdf)(Y=y) = (log o cdf o g^{-1})(y) Programmatically: distribution.log_cdf(bijector.inverse(x))

• and similarly for: cdf, prob, log_survival_function, survival_function.

Kullback-Leibler divergence is also well defined for TransformedDistribution instances that have matching bijectors. Bijector matching is performed via the Bijector.eq method, e.g., td1.bijector == td2.bijector, as part of the KL calculation. If the underlying bijectors do not match, a NotImplementedError is raised when calling kl_divergence. This is the same behavior as calling kl_divergence when two distributions do not have a registered KL divergence.

A simple example constructing a Log-Normal distribution from a Normal distribution:

tfd = tfp.distributions
tfb = tfp.bijectors
log_normal = tfd.TransformedDistribution(
  distribution=tfd.Normal(loc=0., scale=1.),
  bijector=tfb.Exp(),
  name='LogNormalTransformedDistribution')

A LogNormal made from callables:

tfd = tfp.distributions
tfb = tfp.bijectors
log_normal = tfd.TransformedDistribution(
  distribution=tfd.Normal(loc=0., scale=1.),
  bijector=tfb.Inline(
    forward_fn=tf.exp,
    inverse_fn=tf.log,
    inverse_log_det_jacobian_fn=(
      lambda y: -tf.reduce_sum(tf.log(y), axis=-1)),
  name='LogNormalTransformedDistribution')

Another example constructing a Normal from a StandardNormal:

tfd = tfp.distributions
tfb = tfp.bijectors
normal = tfd.TransformedDistribution(
  distribution=tfd.Normal(loc=0., scale=1.),
  bijector=tfb.Shift(shift=-1.)(tfb.Scale(scale=2.)),
  name='NormalTransformedDistribution')

A TransformedDistribution's batch_shape is the same as that of the base distribution, and its event_shape is the forward_event_shape of the bijector applied to the event_shape of the base distribution.

tfd.Sample or tfd.Independent may be used to add extra IID dimensions to the event_shape of the base distribution before the bijector operates on it. The following example demonstrates how to construct a multivariate Normal as a TransformedDistribution, by adding a rank-1 IID dimension to the event_shape of a standard Normal and applying tfb.ScaleMatvecTriL.

tfd = tfp.distributions
tfb = tfp.bijectors
# We will create two MVNs with batch_shape = event_shape = 2.
mean = [[-1., 0],      # batch:0
        [0., 1]]       # batch:1
chol_cov = [[[1., 0],
             [0, 1]],  # batch:0
            [[1, 0],
             [2, 2]]]  # batch:1
mvn1 = tfd.TransformedDistribution(
    distribution=tfd.Sample(
        tfd.Normal(loc=[0., 0], scale=1.),  # base_dist.batch_shape == [2]
        sample_shape=[2])                   # base_dist.event_shape == [2]
    bijector=tfb.Shift(shift=mean)(tfb.ScaleMatvecTriL(scale_tril=chol_cov)))
mvn2 = ds.MultivariateNormalTriL(loc=mean, scale_tril=chol_cov)
# mvn1.log_prob(x) == mvn2.log_prob(x)

a = tf.Module()
b = tf.Module()
c = tf.Module()
a.b = b
b.c = c
list(a.submodules) == [b, c]
True
list(b.submodules) == [c]
True
list(c.submodules) == []
True

Methods

batch_shape_tensor

View source

batch_shape_tensor(
    name='batch_shape_tensor'
)

Shape of a single sample from a single event index as a 1-D Tensor.

The batch dimensions are indexes into independent, non-identical parameterizations of this distribution.

cdf

View source

cdf(
    value, name='cdf', **kwargs
)

Cumulative distribution function.

Given random variable X, the cumulative distribution function cdf is:

cdf(x) := P[X <= x]

copy

View source

copy(
    **override_parameters_kwargs
)

Creates a deep copy of the distribution.

covariance

View source

covariance(
    name='covariance', **kwargs
)

Covariance.

Covariance is (possibly) defined only for non-scalar-event distributions.

For example, for a length-k, vector-valued distribution, it is calculated as,

Cov[i, j] = Covariance(X_i, X_j) = E[(X_i - E[X_i]) (X_j - E[X_j])]

where Cov is a (batch of) k x k matrix, 0 <= (i, j) < k, and E denotes expectation.

Alternatively, for non-vector, multivariate distributions (e.g., matrix-valued, Wishart), Covariance shall return a (batch of) matrices under some vectorization of the events, i.e.,

Cov[i, j] = Covariance(Vec(X)_i, Vec(X)_j) = [as above]

where Cov is a (batch of) k' x k' matrices, 0 <= (i, j) < k' = reduce_prod(event_shape), and Vec is some function mapping indices of this distribution's event dimensions to indices of a length-k' vector.

cross_entropy

View source

cross_entropy(
    other, name='cross_entropy'
)

Computes the (Shannon) cross entropy.

Denote this distribution (self) by P and the other distribution by Q. Assuming P, Q are absolutely continuous with respect to one another and permit densities p(x) dr(x) and q(x) dr(x), (Shannon) cross entropy is defined as:

H[P, Q] = E_p[-log q(X)] = -int_F p(x) log q(x) dr(x)

where F denotes the support of the random variable X ~ P.

other types with built-in registrations: Chi, ExpInverseGamma, GeneralizedExtremeValue, Gumbel, JohnsonSU, Kumaraswamy, LambertWDistribution, LambertWNormal, LogLogistic, LogNormal, LogitNormal, Moyal, MultivariateNormalDiag, MultivariateNormalDiagPlusLowRank, MultivariateNormalFullCovariance, MultivariateNormalLinearOperator, MultivariateNormalTriL, RelaxedOneHotCategorical, SinhArcsinh, TransformedDistribution, VectorExponentialDiag, Weibull

entropy

View source

entropy(
    name='entropy', **kwargs
)

Shannon entropy in nats.

event_shape_tensor

View source

event_shape_tensor(
    name='event_shape_tensor'
)

Shape of a single sample from a single batch as a 1-D int32 Tensor.

experimental_default_event_space_bijector

View source

experimental_default_event_space_bijector(
    *args, **kwargs
)

Bijector mapping the reals (R**n) to the event space of the distribution.

Distributions with continuous support may implement _default_event_space_bijector which returns a subclass of tfp.bijectors.Bijector that maps R**n to the distribution's event space. For example, the default bijector for the Beta distribution is tfp.bijectors.Sigmoid(), which maps the real line to [0, 1], the support of the Beta distribution. The default bijector for the CholeskyLKJ distribution is tfp.bijectors.CorrelationCholesky, which maps R^(k * (k-1) // 2) to the submanifold of k x k lower triangular matrices with ones along the diagonal.

The purpose of experimental_default_event_space_bijector is to enable gradient descent in an unconstrained space for Variational Inference and Hamiltonian Monte Carlo methods. Some effort has been made to choose bijectors such that the tails of the distribution in the unconstrained space are between Gaussian and Exponential.

For distributions with discrete event space, or for which TFP currently lacks a suitable bijector, this function returns None.

is_scalar_batch

View source

is_scalar_batch(
    name='is_scalar_batch'
)

Indicates that batch_shape == [].

is_scalar_event

View source

is_scalar_event(
    name='is_scalar_event'
)

Indicates that event_shape == [].

kl_divergence

View source

kl_divergence(
    other, name='kl_divergence'
)

Computes the Kullback--Leibler divergence.

Denote this distribution (self) by p and the other distribution by q. Assuming p, q are absolutely continuous with respect to reference measure r, the KL divergence is defined as:

KL[p, q] = E_p[log(p(X)/q(X))]
         = -int_F p(x) log q(x) dr(x) + int_F p(x) log p(x) dr(x)
         = H[p, q] - H[p]

where F denotes the support of the random variable X ~ p, H[., .] denotes (Shannon) cross entropy, and H[.] denotes (Shannon) entropy.

other types with built-in registrations: Chi, ExpInverseGamma, GeneralizedExtremeValue, Gumbel, JohnsonSU, Kumaraswamy, LambertWDistribution, LambertWNormal, LogLogistic, LogNormal, LogitNormal, Moyal, MultivariateNormalDiag, MultivariateNormalDiagPlusLowRank, MultivariateNormalFullCovariance, MultivariateNormalLinearOperator, MultivariateNormalTriL, RelaxedOneHotCategorical, SinhArcsinh, TransformedDistribution, VectorExponentialDiag, Weibull

log_cdf

View source

log_cdf(
    value, name='log_cdf', **kwargs
)

Log cumulative distribution function.

Given random variable X, the cumulative distribution function cdf is:

log_cdf(x) := Log[ P[X <= x] ]

Often, a numerical approximation can be used for log_cdf(x) that yields a more accurate answer than simply taking the logarithm of the cdf when x << -1.

log_prob

View source

log_prob(
    value, name='log_prob', **kwargs
)

Log probability density/mass function.

log_survival_function

View source

log_survival_function(
    value, name='log_survival_function', **kwargs
)

Log survival function.

Given random variable X, the survival function is defined:

log_survival_function(x) = Log[ P[X > x] ]
                         = Log[ 1 - P[X <= x] ]
                         = Log[ 1 - cdf(x) ]

Typically, different numerical approximations can be used for the log survival function, which are more accurate than 1 - cdf(x) when x >> 1.

mean

View source

mean(
    name='mean', **kwargs
)

Mean.

mode

View source

mode(
    name='mode', **kwargs
)

Mode.

param_shapes

View source

@classmethod
param_shapes(
    sample_shape, name='DistributionParamShapes'
)

Shapes of parameters given the desired shape of a call to sample(). (deprecated)

This is a class method that describes what key/value arguments are required to instantiate the given Distribution so that a particular shape is returned for that instance's call to sample().

Subclasses should override class method _param_shapes.

param_static_shapes

View source

@classmethod
param_static_shapes(
    sample_shape
)

param_shapes with static (i.e. TensorShape) shapes. (deprecated)

This is a class method that describes what key/value arguments are required to instantiate the given Distribution so that a particular shape is returned for that instance's call to sample(). Assumes that the sample's shape is known statically.

Subclasses should override class method _param_shapes to return constant-valued tensors when constant values are fed.

parameter_properties

View source

@classmethod
parameter_properties(
    dtype=tf.float32, num_classes=None
)

Returns a dict mapping constructor arg names to property annotations.

This dict should include an entry for each of the distribution's Tensor-valued constructor arguments.

Distribution subclasses are not required to implement _parameter_properties, so this method may raise NotImplementedError. Providing a _parameter_properties implementation enables several advanced features, including:

• Distribution batch slicing (sliced_distribution = distribution[i:j]).
• Automatic inference of _batch_shape and _batch_shape_tensor, which must otherwise be computed explicitly.
• Automatic instantiation of the distribution within TFP's internal property tests.
• Automatic construction of 'trainable' instances of the distribution using appropriate bijectors to avoid violating parameter constraints. This enables the distribution family to be used easily as a surrogate posterior in variational inference.

In the future, parameter property annotations may enable additional functionality; for example, returning Distribution instances from tf.vectorized_map.

prob

View source

prob(
    value, name='prob', **kwargs
)

Probability density/mass function.

quantile

View source

quantile(
    value, name='quantile', **kwargs
)

Quantile function. Aka 'inverse cdf' or 'percent point function'.

Given random variable X and p in [0, 1], the quantile is:

quantile(p) := x such that P[X <= x] == p

sample

View source

sample(
    sample_shape=(), seed=None, name='sample', **kwargs
)

Generate samples of the specified shape.

Note that a call to sample() without arguments will generate a single sample.

stddev

View source

stddev(
    name='stddev', **kwargs
)

Standard deviation.

Standard deviation is defined as,

stddev = E[(X - E[X])**2]**0.5

where X is the random variable associated with this distribution, E denotes expectation, and stddev.shape = batch_shape + event_shape.

survival_function

View source

survival_function(
    value, name='survival_function', **kwargs
)

Survival function.

Given random variable X, the survival function is defined:

survival_function(x) = P[X > x]
                     = 1 - P[X <= x]
                     = 1 - cdf(x).

unnormalized_log_prob

View source

unnormalized_log_prob(
    value, name='unnormalized_log_prob', **kwargs
)

Potentially unnormalized log probability density/mass function.

This function is similar to log_prob, but does not require that the return value be normalized. (Normalization here refers to the total integral of probability being one, as it should be by definition for any probability distribution.) This is useful, for example, for distributions where the normalization constant is difficult or expensive to compute. By default, this simply calls log_prob.

variance

View source

variance(
    name='variance', **kwargs
)

Variance.

Variance is defined as,

Var = E[(X - E[X])**2]

where X is the random variable associated with this distribution, E denotes expectation, and Var.shape = batch_shape + event_shape.

with_name_scope

@classmethod
with_name_scope(
    method
)

Decorator to automatically enter the module name scope.

class MyModule(tf.Module):
  @tf.Module.with_name_scope
  def __call__(self, x):
    if not hasattr(self, 'w'):
      self.w = tf.Variable(tf.random.normal([x.shape[1], 3]))
    return tf.matmul(x, self.w)

Using the above module would produce tf.Variables and tf.Tensors whose names included the module name:

mod = MyModule()
mod(tf.ones([1, 2]))
<tf.Tensor: shape=(1, 3), dtype=float32, numpy=..., dtype=float32)>
mod.w
<tf.Variable 'my_module/Variable:0' shape=(2, 3) dtype=float32,
numpy=..., dtype=float32)>

__getitem__

View source

__getitem__(
    slices
)

Slices the batch axes of this distribution, returning a new instance.

b = tfd.Bernoulli(logits=tf.zeros([3, 5, 7, 9]))
b.batch_shape  # => [3, 5, 7, 9]
b2 = b[:, tf.newaxis, ..., -2:, 1::2]
b2.batch_shape  # => [3, 1, 5, 2, 4]

x = tf.random.normal([5, 3, 2, 2])
cov = tf.matmul(x, x, transpose_b=True)
chol = tf.linalg.cholesky(cov)
loc = tf.random.normal([4, 1, 3, 1])
mvn = tfd.MultivariateNormalTriL(loc, chol)
mvn.batch_shape  # => [4, 5, 3]
mvn.event_shape  # => [2]
mvn2 = mvn[:, 3:, ..., ::-1, tf.newaxis]
mvn2.batch_shape  # => [4, 2, 3, 1]
mvn2.event_shape  # => [2]

__iter__

View source

__iter__()