Distributions of Interest

The distribution of interest lets us specify the set of samples over which we want our explanations to be faithful. In some cases, we may want to explain the model’s behavior on a particular record, whereas other times we may be interested in a more general behavior over a distribution of samples.

DoI

Bases: AbstractBaseClass

Interface for distributions of interest. The Distribution of Interest (DoI) specifies the samples over which an attribution method is aggregated.

Source code in trulens_explain/trulens/nn/distributions.py

class DoI(AbstractBaseClass):
    """
    Interface for distributions of interest. The *Distribution of Interest* 
    (DoI) specifies the samples over which an attribution method is 
    aggregated.
    """

    def __init__(self, cut: Cut = None):
        """"Initialize DoI

        Parameters:
            cut (Cut, optional): 
                The Cut in which the DoI will be applied. If `None`, the DoI will be
                applied to the input. otherwise, the distribution should be applied
                to the latent space defined by the cut. 
        """
        self._cut = cut

    def __str__(self):
        return render_object(self, ['_cut'])

    def _wrap_public_call(
        self, z: Inputs[TensorLike], *, model_inputs: ModelInputs
    ) -> Inputs[Uniform[TensorLike]]:
        """Same as __call__ but input and output types are more specific and
        less permissive. Formats the inputs for special cases that might be more
        convenient for the user's __call__ implementation and formats its return
        back to the consistent type."""

        z: Inputs[TensorLike] = om_of_many(z)

        if accepts_model_inputs(self.__call__):
            ret = self.__call__(z, model_inputs=model_inputs)
        else:
            ret = self.__call__(z)
        # Wrap the public doi generator with appropriate type aliases.
        if isinstance(ret, DATA_CONTAINER_TYPE):
            if isinstance(ret[0], DATA_CONTAINER_TYPE):
                ret = Inputs(Uniform(x) for x in ret)
            else:
                ret = Uniform(ret)

            ret: Inputs[Uniform[TensorLike]] = many_of_om(
                ret, innertype=Uniform
            )
        else:
            ret: ArgsLike = [ret]
        return ret

    @abstractmethod
    def __call__(
        self,
        z: OM[Inputs, TensorLike],
        *,
        model_inputs: Optional[ModelInputs] = None
    ) -> OM[Inputs, Uniform[TensorLike]]:
        """
        Computes the distribution of interest from an initial point. If z:
        TensorLike is given, we assume there is only 1 input to the DoI layer. If
        z: List[TensorLike] is given, it provides all of the inputs to the DoI
        layer. 

        Either way, we always return List[List[TensorLike]] (alias
        Inputs[Uniform[TensorLike]]) with outer list spanning layer inputs, and
        inner list spanning a distribution's instance.

        Parameters:
            z:
                Input point from which the distribution is derived. If
                list/tuple, the point is defined by multiple tensors.
            model_inputs:
                Optional wrapped model input arguments that produce value z at
                cut.

        Returns:
            List of points which are all assigned equal probability mass in the
            distribution of interest, i.e., the distribution of interest is a
            discrete, uniform distribution over the list of returned points. If
            z is multi-input, returns a distribution for each input.
        """
        raise NotImplementedError

    # @property
    def cut(self) -> Cut:
        """
        Returns:
            The Cut in which the DoI will be applied. If `None`, the DoI will be
            applied to the input. otherwise, the distribution should be applied
            to the latent space defined by the cut. 
        """
        return self._cut

    def _wrap_public_get_activation_multiplier(
        self, activation: Inputs[TensorLike], *, model_inputs: ModelInputs
    ) -> Inputs[TensorLike]:
        """Same as get_activation_multiplier but without "one-or-more". """

        activations: OM[Inputs, TensorLike] = om_of_many(activation)

        # get_activation_multiplier is public
        if accepts_model_inputs(self.get_activation_multiplier):
            ret: OM[Inputs, TensorLike] = self.get_activation_multiplier(
                activations, model_inputs=model_inputs
            )
        else:
            ret: OM[Inputs,
                    TensorLike] = self.get_activation_multiplier(activations)

        ret: Inputs[TensorLike] = many_of_om(ret)

        return ret

    def get_activation_multiplier(
        self,
        activation: OM[Inputs, TensorLike],
        *,
        model_inputs: Optional[ModelInputs] = None
    ) -> OM[Inputs, TensorLike]:
        """
        Returns a term to multiply the gradient by to convert from "*influence
        space*" to "*attribution space*". Conceptually, "influence space"
        corresponds to the potential effect of a slight increase in each
        feature, while "attribution space" corresponds to an approximation of
        the net marginal contribution to the quantity of interest of each
        feature.

        Parameters:
            activation:
                The activation of the layer the DoI is applied to. DoI may be
                multi-input in which case activation will be a list.
            model_inputs:
                Optional wrapped model input arguments that produce activation
                at cut.

        Returns:
            An array with the same shape as ``activation`` that will be
            multiplied by the gradient to obtain the attribution. The default
            implementation of this method simply returns ``activation``. If
            activation is multi-input, returns one multiplier for each.
        """
        return om_of_many(activation)

    def _assert_cut_contains_only_one_tensor(self, x):
        if isinstance(x, DATA_CONTAINER_TYPE) and len(x) == 1:
            x = x[0]
        if isinstance(x, MAP_CONTAINER_TYPE) and len(x) == 1:
            x = list(x.values())[0]

        if isinstance(x, list):
            raise DoiCutSupportError(
                '\n\n'
                'Cut provided to distribution of interest was comprised of '
                'multiple tensors, but `{}` is only defined for cuts comprised '
                'of a single tensor (received a list of {} tensors).\n'
                '\n'
                'Either (1) select a slice where the `to_cut` corresponds to a '
                'single tensor, or (2) implement/use a `DoI` object that '
                'supports lists of tensors, i.e., where the parameter, `z`, to '
                '`__call__` is expected/allowed to be a list of {} tensors.'.
                format(self.__class__.__name__, len(x), len(x))
            )

        elif not (isinstance(x, np.ndarray) or get_backend().is_tensor(x)):
            raise ValueError(
                '`{}` expected to receive an instance of `Tensor` or '
                '`np.ndarray`, but received an instance of {}'.format(
                    self.__class__.__name__, type(x)
                )
            )

__call__(z, *, model_inputs=None) abstractmethod

Computes the distribution of interest from an initial point. If z: TensorLike is given, we assume there is only 1 input to the DoI layer. If z: List[TensorLike] is given, it provides all of the inputs to the DoI layer.

Either way, we always return List[List[TensorLike]] (alias Inputs[Uniform[TensorLike]]) with outer list spanning layer inputs, and inner list spanning a distribution's instance.

Parameters:

Name	Type	Description	Default
z	OM[Inputs, TensorLike]	Input point from which the distribution is derived. If list/tuple, the point is defined by multiple tensors.	required
model_inputs	Optional[ModelInputs]	Optional wrapped model input arguments that produce value z at cut.	None

Returns:

Type	Description
OM[Inputs, Uniform[TensorLike]]	List of points which are all assigned equal probability mass in the
OM[Inputs, Uniform[TensorLike]]	distribution of interest, i.e., the distribution of interest is a
OM[Inputs, Uniform[TensorLike]]	discrete, uniform distribution over the list of returned points. If
OM[Inputs, Uniform[TensorLike]]	z is multi-input, returns a distribution for each input.

Source code in trulens_explain/trulens/nn/distributions.py

@abstractmethod
def __call__(
    self,
    z: OM[Inputs, TensorLike],
    *,
    model_inputs: Optional[ModelInputs] = None
) -> OM[Inputs, Uniform[TensorLike]]:
    """
    Computes the distribution of interest from an initial point. If z:
    TensorLike is given, we assume there is only 1 input to the DoI layer. If
    z: List[TensorLike] is given, it provides all of the inputs to the DoI
    layer. 

    Either way, we always return List[List[TensorLike]] (alias
    Inputs[Uniform[TensorLike]]) with outer list spanning layer inputs, and
    inner list spanning a distribution's instance.

    Parameters:
        z:
            Input point from which the distribution is derived. If
            list/tuple, the point is defined by multiple tensors.
        model_inputs:
            Optional wrapped model input arguments that produce value z at
            cut.

    Returns:
        List of points which are all assigned equal probability mass in the
        distribution of interest, i.e., the distribution of interest is a
        discrete, uniform distribution over the list of returned points. If
        z is multi-input, returns a distribution for each input.
    """
    raise NotImplementedError

__init__(cut=None)

"Initialize DoI

Parameters:

Name	Type	Description	Default
cut	Cut	The Cut in which the DoI will be applied. If None, the DoI will be applied to the input. otherwise, the distribution should be applied to the latent space defined by the cut.	None

Source code in trulens_explain/trulens/nn/distributions.py

def __init__(self, cut: Cut = None):
    """"Initialize DoI

    Parameters:
        cut (Cut, optional): 
            The Cut in which the DoI will be applied. If `None`, the DoI will be
            applied to the input. otherwise, the distribution should be applied
            to the latent space defined by the cut. 
    """
    self._cut = cut

cut()

Returns:

Type	Description
Cut	The Cut in which the DoI will be applied. If None, the DoI will be
Cut	applied to the input. otherwise, the distribution should be applied
Cut	to the latent space defined by the cut.

Source code in trulens_explain/trulens/nn/distributions.py

def cut(self) -> Cut:
    """
    Returns:
        The Cut in which the DoI will be applied. If `None`, the DoI will be
        applied to the input. otherwise, the distribution should be applied
        to the latent space defined by the cut. 
    """
    return self._cut

get_activation_multiplier(activation, *, model_inputs=None)

Returns a term to multiply the gradient by to convert from "influence space" to "attribution space". Conceptually, "influence space" corresponds to the potential effect of a slight increase in each feature, while "attribution space" corresponds to an approximation of the net marginal contribution to the quantity of interest of each feature.

Parameters:

Name	Type	Description	Default
activation	OM[Inputs, TensorLike]	The activation of the layer the DoI is applied to. DoI may be multi-input in which case activation will be a list.	required
model_inputs	Optional[ModelInputs]	Optional wrapped model input arguments that produce activation at cut.	None

Returns:

Type	Description
OM[Inputs, TensorLike]	An array with the same shape as activation that will be
OM[Inputs, TensorLike]	multiplied by the gradient to obtain the attribution. The default
OM[Inputs, TensorLike]	implementation of this method simply returns activation. If
OM[Inputs, TensorLike]	activation is multi-input, returns one multiplier for each.

Source code in trulens_explain/trulens/nn/distributions.py

def get_activation_multiplier(
    self,
    activation: OM[Inputs, TensorLike],
    *,
    model_inputs: Optional[ModelInputs] = None
) -> OM[Inputs, TensorLike]:
    """
    Returns a term to multiply the gradient by to convert from "*influence
    space*" to "*attribution space*". Conceptually, "influence space"
    corresponds to the potential effect of a slight increase in each
    feature, while "attribution space" corresponds to an approximation of
    the net marginal contribution to the quantity of interest of each
    feature.

    Parameters:
        activation:
            The activation of the layer the DoI is applied to. DoI may be
            multi-input in which case activation will be a list.
        model_inputs:
            Optional wrapped model input arguments that produce activation
            at cut.

    Returns:
        An array with the same shape as ``activation`` that will be
        multiplied by the gradient to obtain the attribution. The default
        implementation of this method simply returns ``activation``. If
        activation is multi-input, returns one multiplier for each.
    """
    return om_of_many(activation)

DoiCutSupportError

Bases: ValueError

Exception raised if the distribution of interest is called on a cut whose output is not supported by the distribution of interest.

Source code in trulens_explain/trulens/nn/distributions.py

class DoiCutSupportError(ValueError):
    """
    Exception raised if the distribution of interest is called on a cut whose
    output is not supported by the distribution of interest.
    """
    pass

GaussianDoi

Bases: DoI

Distribution representing a Gaussian ball around the point. Used by Smooth Gradients.

Source code in trulens_explain/trulens/nn/distributions.py

class GaussianDoi(DoI):
    """
    Distribution representing a Gaussian ball around the point. Used by Smooth
    Gradients.
    """

    def __init__(self, var: float, resolution: int, cut: Cut = None):
        """
        Parameters:
            var:
                The variance of the Gaussian noise to be added around the point.

            resolution:
                Number of samples returned by each call to this DoI.
            cut (Cut, optional): 
                The Cut in which the DoI will be applied. If `None`, the DoI will be
                applied to the input. otherwise, the distribution should be applied
                to the latent space defined by the cut. 
        """
        super(GaussianDoi, self).__init__(cut)
        self._var = var
        self._resolution = resolution

    def __str__(self):
        return render_object(self, ['_cut', '_var', '_resolution'])

    def __call__(self, z: OM[Inputs,
                             TensorLike]) -> OM[Inputs, Uniform[TensorLike]]:
        # Public interface.

        B = get_backend()
        self._assert_cut_contains_only_one_tensor(z)

        def gauss_of_input(z: TensorLike) -> Uniform[TensorLike]:
            # TODO: make a pytorch backend with the same interface to use in places like these.

            if B.is_tensor(z):
                # Tensor implementation.
                return [
                    z + B.random_normal_like(z, var=self._var)
                    for _ in range(self._resolution)
                ]  # Uniform

            else:
                # Array implementation.
                return [
                    z + np.random.normal(0., np.sqrt(self._var), z.shape)
                    for _ in range(self._resolution)
                ]  # Uniform

        z: Inputs[TensorLike] = many_of_om(z)

        return om_of_many(nested_map(z, gauss_of_input))

__init__(var, resolution, cut=None)

Parameters:

Name	Type	Description	Default
var	float	The variance of the Gaussian noise to be added around the point.	required
resolution	int	Number of samples returned by each call to this DoI.	required
cut	Cut	The Cut in which the DoI will be applied. If None, the DoI will be applied to the input. otherwise, the distribution should be applied to the latent space defined by the cut.	None

Source code in trulens_explain/trulens/nn/distributions.py

def __init__(self, var: float, resolution: int, cut: Cut = None):
    """
    Parameters:
        var:
            The variance of the Gaussian noise to be added around the point.

        resolution:
            Number of samples returned by each call to this DoI.
        cut (Cut, optional): 
            The Cut in which the DoI will be applied. If `None`, the DoI will be
            applied to the input. otherwise, the distribution should be applied
            to the latent space defined by the cut. 
    """
    super(GaussianDoi, self).__init__(cut)
    self._var = var
    self._resolution = resolution

LinearDoi

Bases: DoI

Distribution representing the linear interpolation between a baseline and the given point. Used by Integrated Gradients.

Source code in trulens_explain/trulens/nn/distributions.py

class LinearDoi(DoI):
    """
    Distribution representing the linear interpolation between a baseline and 
    the given point. Used by Integrated Gradients.
    """

    def __init__(
        self,
        baseline: BaselineLike = None,
        resolution: int = 10,
        *,
        cut: Cut = None,
    ):
        """
        The DoI for point, `z`, will be a uniform distribution over the points
        on the line segment connecting `z` to `baseline`, approximated by a
        sample of `resolution` points equally spaced along this segment.

        Parameters:
            cut (Cut, optional, from DoI): 
                The Cut in which the DoI will be applied. If `None`, the DoI
                will be applied to the input. otherwise, the distribution should
                be applied to the latent space defined by the cut. 
            baseline (BaselineLike, optional):
                The baseline to interpolate from. Must be same shape as the
                space the distribution acts over, i.e., the shape of the points,
                `z`, eventually passed to `__call__`. If `cut` is `None`, this
                must be the same shape as the input, otherwise this must be the
                same shape as the latent space defined by the cut. If `None` is
                given, `baseline` will be the zero vector in the appropriate
                shape. If the baseline is callable, it is expected to return the
                `baseline`, given `z` and optional model arguments.
            resolution (int):
                Number of points returned by each call to this DoI. A higher
                resolution is more computationally expensive, but gives a better
                approximation of the DoI this object mathematically represents.
        """
        super(LinearDoi, self).__init__(cut)
        self._baseline = baseline
        self._resolution = resolution

    @property
    def baseline(self) -> BaselineLike:
        return self._baseline

    @property
    def resolution(self) -> int:
        return self._resolution

    def __str__(self):
        return render_object(self, ['_cut', '_baseline', '_resolution'])

    def __call__(
        self,
        z: OM[Inputs, TensorLike],
        *,
        model_inputs: Optional[ModelInputs] = None
    ) -> OM[Inputs, Uniform[TensorLike]]:

        self._assert_cut_contains_only_one_tensor(z)

        z: Inputs[TensorLike] = many_of_om(z)

        baseline = self._compute_baseline(z, model_inputs=model_inputs)

        r = 1. if self._resolution == 1 else self._resolution - 1.
        zipped = nested_zip(z, baseline)

        def zipped_interpolate(zipped_z_baseline):
            """interpolates zipped elements

            Args:
                zipped_z_baseline: A tuple expecting the first element to be the z_val, and second to be the baseline.

            Returns:
                a list of interpolations from z to baseline
            """
            z_ = zipped_z_baseline[0]
            b_ = zipped_z_baseline[1]
            return [ # Uniform
                (1. - i / r) * z_ + i / r * b_
                for i in range(self._resolution)
            ]

        ret = om_of_many(
            nested_map(
                zipped, zipped_interpolate, check_accessor=lambda x: x[0]
            )
        )

        return ret

    def get_activation_multiplier(
        self,
        activation: OM[Inputs, TensorLike],
        *,
        model_inputs: Optional[ModelInputs] = None
    ) -> Inputs[TensorLike]:
        """
        Returns a term to multiply the gradient by to convert from "*influence 
        space*" to "*attribution space*". Conceptually, "influence space"
        corresponds to the potential effect of a slight increase in each 
        feature, while "attribution space" corresponds to an approximation of
        the net marginal contribution to the quantity of interest of each 
        feature.

        Parameters:
            activation:
                The activation of the layer the DoI is applied to.

        Returns:
            The activation adjusted by the baseline passed to the constructor.
        """

        activation: Inputs[TensorLike] = many_of_om(activation)

        baseline: Inputs[TensorLike] = self._compute_baseline(
            activation, model_inputs=model_inputs
        )

        if baseline is None:
            return activation

        zipped = nested_zip(activation, baseline)

        def zipped_subtract(zipped_activation_baseline):
            """subtracts zipped elements

            Args:
                zipped_activation_baseline: A tuple expecting the first element to be the activation, and second to be the baseline.

            Returns:
                a subtraction of activation and baseline
            """
            activation = zipped_activation_baseline[0]
            baseline = zipped_activation_baseline[1]
            return activation - baseline

        ret = nested_map(zipped, zipped_subtract, check_accessor=lambda x: x[0])
        return ret

    def _compute_baseline(
        self,
        z: Inputs[TensorLike],
        *,
        model_inputs: Optional[ModelInputs] = None
    ) -> Inputs[TensorLike]:

        B = get_backend()

        _baseline: BaselineLike = self.baseline  # user-provided

        if isinstance(_baseline, Callable):
            if accepts_model_inputs(_baseline):
                _baseline: OM[Inputs, TensorLike] = many_of_om(
                    _baseline(om_of_many(z), model_inputs=model_inputs)
                )
            else:
                _baseline: OM[Inputs, TensorLike] = many_of_om(
                    _baseline(om_of_many(z))
                )

        else:
            _baseline: OM[Inputs, TensorLike]

        if _baseline is None:
            _baseline: Inputs[TensorLike] = nested_map(z, B.zeros_like)
        else:
            _baseline: Inputs[TensorLike] = many_of_om(_baseline)
            # Came from user; could have been single or multiple inputs.
        tensor_wrapper = TensorAKs(args=z)
        # Cast to either Tensor or numpy.ndarray to match what was given in z.
        return nested_cast(
            backend=B,
            args=_baseline,
            astype=type(tensor_wrapper.first_batchable(B))
        )

__init__(baseline=None, resolution=10, *, cut=None)

The DoI for point, z, will be a uniform distribution over the points on the line segment connecting z to baseline, approximated by a sample of resolution points equally spaced along this segment.

Parameters:

Name	Type	Description	Default
cut	Cut, optional, from DoI	The Cut in which the DoI will be applied. If None, the DoI will be applied to the input. otherwise, the distribution should be applied to the latent space defined by the cut.	None
baseline	BaselineLike	The baseline to interpolate from. Must be same shape as the space the distribution acts over, i.e., the shape of the points, z, eventually passed to __call__. If cut is None, this must be the same shape as the input, otherwise this must be the same shape as the latent space defined by the cut. If None is given, baseline will be the zero vector in the appropriate shape. If the baseline is callable, it is expected to return the baseline, given z and optional model arguments.	None
resolution	int	Number of points returned by each call to this DoI. A higher resolution is more computationally expensive, but gives a better approximation of the DoI this object mathematically represents.	10

Source code in trulens_explain/trulens/nn/distributions.py

def __init__(
    self,
    baseline: BaselineLike = None,
    resolution: int = 10,
    *,
    cut: Cut = None,
):
    """
    The DoI for point, `z`, will be a uniform distribution over the points
    on the line segment connecting `z` to `baseline`, approximated by a
    sample of `resolution` points equally spaced along this segment.

    Parameters:
        cut (Cut, optional, from DoI): 
            The Cut in which the DoI will be applied. If `None`, the DoI
            will be applied to the input. otherwise, the distribution should
            be applied to the latent space defined by the cut. 
        baseline (BaselineLike, optional):
            The baseline to interpolate from. Must be same shape as the
            space the distribution acts over, i.e., the shape of the points,
            `z`, eventually passed to `__call__`. If `cut` is `None`, this
            must be the same shape as the input, otherwise this must be the
            same shape as the latent space defined by the cut. If `None` is
            given, `baseline` will be the zero vector in the appropriate
            shape. If the baseline is callable, it is expected to return the
            `baseline`, given `z` and optional model arguments.
        resolution (int):
            Number of points returned by each call to this DoI. A higher
            resolution is more computationally expensive, but gives a better
            approximation of the DoI this object mathematically represents.
    """
    super(LinearDoi, self).__init__(cut)
    self._baseline = baseline
    self._resolution = resolution

get_activation_multiplier(activation, *, model_inputs=None)

Returns a term to multiply the gradient by to convert from "influence space" to "attribution space". Conceptually, "influence space" corresponds to the potential effect of a slight increase in each feature, while "attribution space" corresponds to an approximation of the net marginal contribution to the quantity of interest of each feature.

Parameters:

Name	Type	Description	Default
activation	OM[Inputs, TensorLike]	The activation of the layer the DoI is applied to.	required

Returns:

Type	Description
Inputs[TensorLike]	The activation adjusted by the baseline passed to the constructor.

Source code in trulens_explain/trulens/nn/distributions.py

def get_activation_multiplier(
    self,
    activation: OM[Inputs, TensorLike],
    *,
    model_inputs: Optional[ModelInputs] = None
) -> Inputs[TensorLike]:
    """
    Returns a term to multiply the gradient by to convert from "*influence 
    space*" to "*attribution space*". Conceptually, "influence space"
    corresponds to the potential effect of a slight increase in each 
    feature, while "attribution space" corresponds to an approximation of
    the net marginal contribution to the quantity of interest of each 
    feature.

    Parameters:
        activation:
            The activation of the layer the DoI is applied to.

    Returns:
        The activation adjusted by the baseline passed to the constructor.
    """

    activation: Inputs[TensorLike] = many_of_om(activation)

    baseline: Inputs[TensorLike] = self._compute_baseline(
        activation, model_inputs=model_inputs
    )

    if baseline is None:
        return activation

    zipped = nested_zip(activation, baseline)

    def zipped_subtract(zipped_activation_baseline):
        """subtracts zipped elements

        Args:
            zipped_activation_baseline: A tuple expecting the first element to be the activation, and second to be the baseline.

        Returns:
            a subtraction of activation and baseline
        """
        activation = zipped_activation_baseline[0]
        baseline = zipped_activation_baseline[1]
        return activation - baseline

    ret = nested_map(zipped, zipped_subtract, check_accessor=lambda x: x[0])
    return ret

PointDoi

Bases: DoI

Distribution that puts all probability mass on a single point.

Source code in trulens_explain/trulens/nn/distributions.py

class PointDoi(DoI):
    """
    Distribution that puts all probability mass on a single point.
    """

    def __init__(self, cut: Cut = None):
        """"Initialize PointDoI

        Parameters:
            cut (Cut, optional): 
                The Cut in which the DoI will be applied. If `None`, the DoI will be
                applied to the input. otherwise, the distribution should be applied
                to the latent space defined by the cut. 
        """
        super(PointDoi, self).__init__(cut)

    def __call__(
        self,
        z: OM[Inputs, TensorLike],
        *,
        model_inputs: Optional[ModelInputs] = None
    ) -> OM[Inputs, Uniform[TensorLike]]:

        z: Inputs[TensorLike] = many_of_om(z)

        return om_of_many(nested_map(z, lambda x: [x]))

__init__(cut=None)

"Initialize PointDoI

Parameters:

Name	Type	Description	Default
cut	Cut	The Cut in which the DoI will be applied. If None, the DoI will be applied to the input. otherwise, the distribution should be applied to the latent space defined by the cut.	None

Source code in trulens_explain/trulens/nn/distributions.py

def __init__(self, cut: Cut = None):
    """"Initialize PointDoI

    Parameters:
        cut (Cut, optional): 
            The Cut in which the DoI will be applied. If `None`, the DoI will be
            applied to the input. otherwise, the distribution should be applied
            to the latent space defined by the cut. 
    """
    super(PointDoi, self).__init__(cut)