scvi.distributions.BetaBinomial

class scvi.distributions.BetaBinomial(total_count, alpha=None, beta=None, mu=None, gamma=None, validate_args=False, eps=1e-08)

BETA Beta binomial distribution.

One of the following parameterizations must be provided:

1. (alpha, beta, total_counts) where alpha and beta are the shape parameters of the beta distribution and total_counts is the number of trials.

2. (mu, gamma, total_counts), which is the one used by methylVI. These parameters respectively control the mean and dispersion of the distribution.

In the (mu, gamma) parameterization, samples from the beta-binomial are generated as follows:

1. \(p_i \sim \textrm{Beta}(\mu, \gamma)\)

2. \(y_i \sim \textrm{Ber}(p_i)\)

3. \(y = \sum_{i}y_i\)

Parameters:

• total_count (Tensor) – Number of trials. Must be a non-negative integer.

• alpha (Optional[Tensor] (default: None)) – As in BetaBinomial, serves as the first shape parameterization of the beta distribution. Must be greater than 0.

• beta (Optional[Tensor] (default: None)) – As in BetaBinomial, serves as the second shape parameterization of the beta distribution. Must be greater than 0.

• mu (Optional[Tensor] (default: None)) – Mean of the distribution. Must be within the interval (0, 1).

• gamma (Optional[Tensor] (default: None)) – Dispersion. Must be within the interval (0, 1).

• validate_args (bool (default: False)) – Raise ValueError if arguments do not match the constraints.

• eps (float (default: 1e-08)) – Numerical stability constant. (See Notes)

Notes

Under the hood we use BetaBinomial to implement the Beta-Binomial distribution. Thus, when the user specifies a (mu, gamma) parameterization, we must convert to the (alpha, beta) parameterization used by the underlying Pyro distribution class. During this process, numerical stability issues sometimes cause alpha or beta to be equal to (exactly) zero. This is not allowed (alpha and beta must be strictly greater than 0), so we clamp these values to be greater than a small constant eps.

Attributes table

approx_log_prob_tol
arg_constraints
batch_shape	Returns the shape over which parameters are batched.
concentration0
concentration1
event_dim	Number of dimensions of individual events.
event_shape	Returns the shape of a single sample (without batching).
has_enumerate_support
has_rsample
mean	Returns the mean of the distribution.
mode	Returns the mode of the distribution.
rv	EXPERIMENTAL Switch to the Random Variable DSL for applying transformations to random variables.
stddev	Returns the standard deviation of the distribution.
support
variance	Returns the variance of the distribution.

Methods table

cdf(value)	Returns the cumulative density/mass function evaluated at value.
conjugate_update(other)	EXPERIMENTAL Creates an updated distribution fusing information from another compatible distribution.
entropy()	Returns entropy of distribution, batched over batch_shape.
enumerate_support([expand])	Returns tensor containing all values supported by a discrete distribution.
expand(batch_shape[, _instance])	Returns a new ExpandedDistribution instance with batch dimensions expanded to batch_shape.
expand_by(sample_shape)	Expands a distribution by adding sample_shape to the left side of its batch_shape.
has_rsample_(value)	Force reparameterized or detached sampling on a single distribution instance.
icdf(value)	Returns the inverse cumulative density/mass function evaluated at value.
independent([reinterpreted_batch_ndims])
infer_shapes(**arg_shapes)	Infers batch_shape and event_shape given shapes of args to __init__().
log_prob(value)	Returns the log of the probability density/mass function evaluated at value.
mask(mask)	Masks a distribution by a boolean or boolean-valued tensor that is broadcastable to the distributions batch_shape .
perplexity()	Returns perplexity of distribution, batched over batch_shape.
reshape([sample_shape, extra_event_dims])
rsample([sample_shape])	Generates a sample_shape shaped reparameterized sample or sample_shape shaped batch of reparameterized samples if the distribution parameters are batched.
sample([sample_shape])	Generates a sample_shape shaped sample or sample_shape shaped batch of samples if the distribution parameters are batched.
sample_n(n)	Generates n samples or n batches of samples if the distribution parameters are batched.
score_parts(x, *args, **kwargs)	Computes ingredients for stochastic gradient estimators of ELBO.
set_default_validate_args(value)	Sets whether validation is enabled or disabled.
shape([sample_shape])	The tensor shape of samples from this distribution.
to_event([reinterpreted_batch_ndims])	Reinterprets the n rightmost dimensions of this distributions batch_shape as event dims, adding them to the left side of event_shape.

Attributes

BetaBinomial.approx_log_prob_tol = 0.0

BetaBinomial.arg_constraints = {'alpha': Optional(GreaterThan(lower_bound=0)), 'beta': Optional(GreaterThan(lower_bound=0)), 'gamma': Optional(OpenInterval(lower_bound=0, upper_bound=1)), 'mu': Optional(OpenInterval(lower_bound=0, upper_bound=1)), 'total_count': IntegerGreaterThan(lower_bound=0)}

BetaBinomial.batch_shape

Returns the shape over which parameters are batched.

BetaBinomial.concentration0

BetaBinomial.concentration1

BetaBinomial.event_dim

Number of dimensions of individual events. :rtype: int

Type:

return

BetaBinomial.event_shape

Returns the shape of a single sample (without batching).

BetaBinomial.has_enumerate_support = True

BetaBinomial.has_rsample = False

BetaBinomial.mean

BetaBinomial.mode

Returns the mode of the distribution.

BetaBinomial.rv

EXPERIMENTAL Switch to the Random Variable DSL for applying transformations to random variables. Supports either chaining operations or arithmetic operator overloading.

Example usage:

# This should be equivalent to an Exponential distribution.
Uniform(0, 1).rv.log().neg().dist

# These two distributions Y1, Y2 should be the same
X = Uniform(0, 1).rv
Y1 = X.mul(4).pow(0.5).sub(1).abs().neg().dist
Y2 = (-abs((4*X)**(0.5) - 1)).dist

Returns:

A :class: ~pyro.contrib.randomvariable.random_variable.RandomVariable object wrapping this distribution.

Return type:

RandomVariable

BetaBinomial.stddev

Returns the standard deviation of the distribution.

BetaBinomial.support = IntegerGreaterThan(lower_bound=0)

BetaBinomial.variance

Methods

BetaBinomial.cdf(value)

Returns the cumulative density/mass function evaluated at value.

Parameters:

value (Tensor)

Return type:

Tensor

BetaBinomial.conjugate_update(other)

EXPERIMENTAL Creates an updated distribution fusing information from another compatible distribution. This is supported by only a few conjugate distributions.

This should satisfy the equation:

fg, log_normalizer = f.conjugate_update(g)
assert f.log_prob(x) + g.log_prob(x) == fg.log_prob(x) + log_normalizer

Note this is equivalent to funsor.ops.add on Funsor distributions, but we return a lazy sum (updated, log_normalizer) because PyTorch distributions must be normalized. Thus conjugate_update() should commute with dist_to_funsor() and tensor_to_funsor()

dist_to_funsor(f) + dist_to_funsor(g)
  == dist_to_funsor(fg) + tensor_to_funsor(log_normalizer)

Parameters:

other – A distribution representing p(data|latent) but normalized over latent rather than data. Here latent is a candidate sample from self and data is a ground observation of unrelated type.

Returns:

a pair (updated,log_normalizer) where updated is an updated distribution of type type(self), and log_normalizer is a Tensor representing the normalization factor.

BetaBinomial.entropy()

Returns entropy of distribution, batched over batch_shape.

Return type:

Tensor

Returns:

Tensor of shape batch_shape.

BetaBinomial.enumerate_support(expand=True)

Returns tensor containing all values supported by a discrete distribution. The result will enumerate over dimension 0, so the shape of the result will be (cardinality,) + batch_shape + event_shape (where event_shape = () for univariate distributions).

Note that this enumerates over all batched tensors in lock-step [[0, 0], [1, 1], …]. With expand=False, enumeration happens along dim 0, but with the remaining batch dimensions being singleton dimensions, [[0], [1], ...

To iterate over the full Cartesian product use itertools.product(m.enumerate_support()).

Parameters:

expand (bool) – whether to expand the support over the batch dims to match the distribution’s batch_shape.

Returns:

Tensor iterating over dimension 0.

BetaBinomial.expand(batch_shape, _instance=None)

Returns a new ExpandedDistribution instance with batch dimensions expanded to batch_shape.

Parameters:

• batch_shape (tuple) – batch shape to expand to.

• _instance (default: None) – unused argument for compatibility with torch.distributions.Distribution.expand()

Returns:

an instance of ExpandedDistribution.

Return type:

ExpandedDistribution

BetaBinomial.expand_by(sample_shape)

Expands a distribution by adding sample_shape to the left side of its batch_shape.

To expand internal dims of self.batch_shape from 1 to something larger, use expand() instead.

Parameters:

sample_shape (torch.Size) – The size of the iid batch to be drawn from the distribution.

Returns:

An expanded version of this distribution.

Return type:

ExpandedDistribution

BetaBinomial.has_rsample_(value)

Force reparameterized or detached sampling on a single distribution instance. This sets the .has_rsample attribute in-place.

This is useful to instruct inference algorithms to avoid reparameterized gradients for variables that discontinuously determine downstream control flow.

Parameters:

value (bool) – Whether samples will be pathwise differentiable.

Returns:

self

Return type:

Distribution

BetaBinomial.icdf(value)

Returns the inverse cumulative density/mass function evaluated at value.

Parameters:

value (Tensor)

Return type:

Tensor

BetaBinomial.independent(reinterpreted_batch_ndims=None)

classmethod BetaBinomial.infer_shapes(**arg_shapes)

Infers batch_shape and event_shape given shapes of args to __init__().

Note

This assumes distribution shape depends only on the shapes of tensor inputs, not in the data contained in those inputs.

Parameters:

**arg_shapes – Keywords mapping name of input arg to torch.Size or tuple representing the sizes of each tensor input.

Returns:

A pair (batch_shape, event_shape) of the shapes of a distribution that would be created with input args of the given shapes.

Return type:

tuple

BetaBinomial.log_prob(value)

Returns the log of the probability density/mass function evaluated at value.

Parameters:

value (Tensor)

BetaBinomial.mask(mask)

Masks a distribution by a boolean or boolean-valued tensor that is broadcastable to the distributions batch_shape .

Parameters:

mask (bool or torch.Tensor) – A boolean or boolean valued tensor.

Returns:

A masked copy of this distribution.

Return type:

MaskedDistribution

BetaBinomial.perplexity()

Returns perplexity of distribution, batched over batch_shape.

Return type:

Tensor

Returns:

Tensor of shape batch_shape.

BetaBinomial.reshape(sample_shape=None, extra_event_dims=None)

BetaBinomial.rsample(sample_shape=())

Generates a sample_shape shaped reparameterized sample or sample_shape shaped batch of reparameterized samples if the distribution parameters are batched.

Return type:

Tensor

BetaBinomial.sample(sample_shape=())

Generates a sample_shape shaped sample or sample_shape shaped batch of samples if the distribution parameters are batched.

BetaBinomial.sample_n(n)

Generates n samples or n batches of samples if the distribution parameters are batched.

Return type:

Tensor

BetaBinomial.score_parts(x, *args, **kwargs)

Computes ingredients for stochastic gradient estimators of ELBO.

The default implementation is correct both for non-reparameterized and for fully reparameterized distributions. Partially reparameterized distributions should override this method to compute correct .score_function and .entropy_term parts.

Setting .has_rsample on a distribution instance will determine whether inference engines like SVI use reparameterized samplers or the score function estimator.

Parameters:

x (torch.Tensor) – A single value or batch of values.

Returns:

A ScoreParts object containing parts of the ELBO estimator.

Return type:

ScoreParts

static BetaBinomial.set_default_validate_args(value)

Sets whether validation is enabled or disabled.

The default behavior mimics Python’s assert statement: validation is on by default, but is disabled if Python is run in optimized mode (via python -O). Validation may be expensive, so you may want to disable it once a model is working.

Parameters:

value (bool) – Whether to enable validation.

Return type:

None

BetaBinomial.shape(sample_shape=())

The tensor shape of samples from this distribution.

Samples are of shape:

d.shape(sample_shape) == sample_shape + d.batch_shape + d.event_shape

Parameters:

sample_shape (torch.Size) – the size of the iid batch to be drawn from the distribution.

Returns:

Tensor shape of samples.

Return type:

torch.Size

BetaBinomial.to_event(reinterpreted_batch_ndims=None)

Reinterprets the n rightmost dimensions of this distributions batch_shape as event dims, adding them to the left side of event_shape.

Example

>>> [d1.batch_shape, d1.event_shape]
[torch.Size([2, 3]), torch.Size([4, 5])]
>>> d2 = d1.to_event(1)
>>> [d2.batch_shape, d2.event_shape]
[torch.Size([2]), torch.Size([3, 4, 5])]
>>> d3 = d1.to_event(2)
>>> [d3.batch_shape, d3.event_shape]
[torch.Size([]), torch.Size([2, 3, 4, 5])]

Parameters:

reinterpreted_batch_ndims (int) – The number of batch dimensions to reinterpret as event dimensions. May be negative to remove dimensions from an pyro.distributions.torch.Independent . If None, convert all dimensions to event dimensions.

Returns:

A reshaped version of this distribution.

Return type:

pyro.distributions.torch.Independent