alibi.explainers.cfrl_base

Constants

TYPE_CHECKING

TYPE_CHECKING: bool = False

bool(x) -> bool

Returns True when the argument x is true, False otherwise. The builtins True and False are the only two instances of the class bool. The class bool is a subclass of the class int, and cannot be subclassed.

DEFAULT_DATA_CFRL

DEFAULT_DATA_CFRL: dict = {'orig': None, 'cf': None, 'target': None, 'condition': None}

DEFAULT_META_CFRL

DEFAULT_META_CFRL: dict = {'name': None, 'type': ['blackbox'], 'explanations': ['local'], 'params': {},...

has_pytorch

has_pytorch: bool = True

bool(x) -> bool

Returns True when the argument x is true, False otherwise. The builtins True and False are the only two instances of the class bool. The class bool is a subclass of the class int, and cannot be subclassed.

has_tensorflow

has_tensorflow: bool = True

bool(x) -> bool

Returns True when the argument x is true, False otherwise. The builtins True and False are the only two instances of the class bool. The class bool is a subclass of the class int, and cannot be subclassed.

logger

logger: logging.Logger = <Logger alibi.explainers.cfrl_base (WARNING)>

Instances of the Logger class represent a single logging channel. A "logging channel" indicates an area of an application. Exactly how an "area" is defined is up to the application developer. Since an application can have any number of areas, logging channels are identified by a unique string. Application areas can be nested (e.g. an area of "input processing" might include sub-areas "read CSV files", "read XLS files" and "read Gnumeric files"). To cater for this natural nesting, channel names are organized into a namespace hierarchy where levels are separated by periods, much like the Java or Python package namespace. So in the instance given above, channel names might be "input" for the upper level, and "input.csv", "input.xls" and "input.gnu" for the sub-levels. There is no arbitrary limit to the depth of nesting.

DEFAULT_BASE_PARAMS

DEFAULT_BASE_PARAMS: dict = {'act_noise': 0.1, 'act_low': -1.0, 'act_high': 1.0, 'replay_buffer_size': 10...

Callback

Inherits from: ABC

Training callback class.

CounterfactualRL

Inherits from: Explainer, FitMixin, ABC, Base

Counterfactual Reinforcement Learning.

Constructor

CounterfactualRL(self, predictor: Callable[[numpy.ndarray], numpy.ndarray], encoder: 'Union[tensorflow.keras.Model, torch.nn.Module]', decoder: 'Union[tensorflow.keras.Model, torch.nn.Module]', coeff_sparsity: float, coeff_consistency: float, latent_dim: Optional[int] = None, backend: str = 'tensorflow', seed: int = 0, **kwargs)

Name	Type	Default	Description
predictor	Callable[[.[<class 'numpy.ndarray'>]], numpy.ndarray]		A callable that takes a numpy array of N data points as inputs and returns N outputs. For classification task, the second dimension of the output should match the number of classes. Thus, the output can be either a soft label distribution or a hard label distribution (i.e. one-hot encoding) without affecting the performance since argmax is applied to the predictor's output.
encoder	Union[tensorflow.keras.Model, torch.nn.Module]		Pretrained encoder network.
decoder	Union[tensorflow.keras.Model, torch.nn.Module]		Pretrained decoder network.
coeff_sparsity	float		Sparsity loss coefficient.
coeff_consistency	float		Consistency loss coefficient.
latent_dim	Optional[int]	None	Auto-encoder latent dimension. Can be omitted if the actor network is user specified.
backend	str	'tensorflow'	Deep learning backend: 'tensorflow'
seed	int	0	Seed for reproducibility. The results are not reproducible for 'tensorflow' backend.
**kwargs			Used to replace any default parameter from :py:data:alibi.explainers.cfrl_base.DEFAULT_BASE_PARAMS.

Methods

explain

explain(X: numpy.ndarray, Y_t: numpy.ndarray, C: Optional[numpy.ndarray] = None, batch_size: int = 100) -> alibi.api.interfaces.Explanation

Name	Type	Default	Description
X	numpy.ndarray		Instances to be explained.
Y_t	numpy.ndarray		Counterfactual targets.
C	Optional[numpy.ndarray]	None	Conditional vectors. If None, it means that no conditioning was used during training (i.e. the conditional_func returns None).
batch_size	int	100	Batch size to be used when generating counterfactuals.

Returns

• Type: alibi.api.interfaces.Explanation

fit

fit(X: numpy.ndarray) -> alibi.api.interfaces.Explainer

Name	Type	Default	Description
X	numpy.ndarray		Training data array.

Returns

• Type: alibi.api.interfaces.Explainer

load

load(path: Union[str, os.PathLike], predictor: typing.Any) -> alibi.api.interfaces.Explainer

Name	Type	Default	Description
path	Union[str, os.PathLike]
predictor	typing.Any

Returns

• Type: alibi.api.interfaces.Explainer

reset_predictor

reset_predictor(predictor: typing.Any) -> None

Name	Type	Default	Description
predictor	typing.Any

Returns

• Type: None

save

save(path: Union[str, os.PathLike]) -> None

Name	Type	Default	Description
path	Union[str, os.PathLike]

Returns

• Type: None

NormalActionNoise

Normal noise generator.

Constructor

NormalActionNoise(self, mu: float, sigma: float) -> None

Name	Type	Default	Description
mu	float		Mean of the normal noise.
sigma	float		Standard deviation of the noise.

Postprocessing

Inherits from: ABC

ReplayBuffer

Circular experience replay buffer for CounterfactualRL (DDPG). When the buffer is filled, then the oldest experience is replaced by the new one (FIFO). The experience batch size is kept constant and inferred when the first batch of data is stored. Allowing flexible batch size can generate tensorflow warning due to the tf.function retracing, which can lead to a drop in performance.

Constructor

ReplayBuffer(self, size: int = 1000) -> None

Name	Type	Default	Description
size	int	1000	Dimension of the buffer in batch size. This that the total memory allocated is proportional with the size x batch_size, where batch_size is inferred from the first array to be stored.

Methods

append

append(X: numpy.ndarray, Y_m: numpy.ndarray, Y_t: numpy.ndarray, Z: numpy.ndarray, Z_cf_tilde: numpy.ndarray, C: Optional[numpy.ndarray], R_tilde: numpy.ndarray, kwargs) -> None

Name	Type	Default	Description
X	numpy.ndarray		Input array.
Y_m	numpy.ndarray		Model's prediction class of X.
Y_t	numpy.ndarray		Counterfactual target class.
Z	numpy.ndarray		Input's embedding.
Z_cf_tilde	numpy.ndarray		Noised counterfactual embedding.
C	Optional[numpy.ndarray]		Conditional array.
R_tilde	numpy.ndarray		Noised counterfactual reward array.
**kwargs			Other arguments. Not used.

Returns

• Type: None

sample

sample() -> Dict[str, Optional[numpy.ndarray]]

Sample a batch of experience form the replay buffer.

Returns

• Type: Dict[str, Optional[numpy.ndarray]]