fusionlab.nn.components.MemoryAugmentedAttention

class fusionlab.nn.components.MemoryAugmentedAttention

Bases: Layer, NNLearner

Memory-Augmented Attention layer that uses a learned memory matrix to enhance temporal representation [1].

This layer maintains a trainable memory of shape \((\text{memory_size}, \text{units})\) and attends over it with the input serving as the query. The resulting context is added back to the input as a residual connection, giving a memory-augmented feature.

\[\mathbf{Z} = \mathbf{X} + \text{MHA}(\mathbf{X}, \mathbf{M}, \mathbf{M})\]

where \(\mathbf{M}\) is the learned memory.

Parameters:

• units (int) – Dimensionality for the memory and the multi-head attention projections.

• memory_size (int) – Number of slots in the learned memory matrix.

• num_heads (int) – Number of attention heads in the multi-head attention.

Notes

The learned memory is a trainable parameter of shape (memory_size, units). It is expanded at each forward pass to match the batch size.

call(`inputs`, training=False)

Forward pass of the memory-augmented attention layer.

get_config()

Returns the configuration for serialization.

from_config(`config`)

Instantiates the layer from the given config dictionary.

Examples

>>> from fusionlab.nn.components import MemoryAugmentedAttention
>>> import tensorflow as tf
>>> # Suppose we have an input of shape (batch_size, time_steps, units)
>>> x = tf.random.normal((32, 10, 64))
>>> # Instantiate with a memory size of 20
>>> maa = MemoryAugmentedAttention(
...     units=64,
...     memory_size=20,
...     num_heads=4
... )
>>> # Forward pass
>>> outputs = maa(x)

See also

CrossAttention

Another specialized attention mechanism focusing on cross-sequence interactions.

HierarchicalAttention

Combines short/long-term sequences with attention.

References

[1]

Graves, A., Wayne, G., & Danihelka, I. (2014). Neural Turing Machines. arXiv preprint arXiv:1410.5401.

__init__(units, memory_size, num_heads)

Parameters:

• units (int)

• memory_size (int)

• num_heads (int)

Methods

__init__(units, memory_size, num_heads)
add_loss(losses, **kwargs)	Add loss tensor(s), potentially dependent on layer inputs.
add_metric(value[, name])	Adds metric tensor to the layer.
add_update(updates)	Add update op(s), potentially dependent on layer inputs.
add_variable(*args, **kwargs)	Deprecated, do NOT use! Alias for add_weight.
add_weight([name, shape, dtype, ...])	Adds a new variable to the layer.
build(input_shape)	Build method that creates the trainable memory matrix of shape (memory_size, units).
build_from_config(config)	Builds the layer's states with the supplied config dict.
call(inputs[, training])	Forward pass of MemoryAugmentedAttention.
compute_mask(inputs[, mask])	Computes an output mask tensor.
compute_output_shape(input_shape)	Computes the output shape of the layer.
compute_output_signature(input_signature)	Compute the output tensor signature of the layer based on the inputs.
count_params()	Count the total number of scalars composing the weights.
finalize_state()	Finalizes the layers state after updating layer weights.
from_config(config)	Creates a new instance from a given config dictionary.
get_build_config()	Returns a dictionary with the layer's input shape.
get_config()	Returns configuration of this layer.
get_input_at(node_index)	Retrieves the input tensor(s) of a layer at a given node.
get_input_mask_at(node_index)	Retrieves the input mask tensor(s) of a layer at a given node.
get_input_shape_at(node_index)	Retrieves the input shape(s) of a layer at a given node.
get_output_at(node_index)	Retrieves the output tensor(s) of a layer at a given node.
get_output_mask_at(node_index)	Retrieves the output mask tensor(s) of a layer at a given node.
get_output_shape_at(node_index)	Retrieves the output shape(s) of a layer at a given node.
get_params([deep])	Get the parameters for this learner.
get_weights()	Returns the current weights of the layer, as NumPy arrays.
help(**kwargs)
load(file_path[, format])	Load the learner's state from a specified file in the desired format.
load_own_variables(store)	Loads the state of the layer.
save([file_path, format, overwrite, ...])	Save the learner's state to a specified file in the desired format.
save_own_variables(store)	Saves the state of the layer.
set_params(**params)	Set the parameters of this learner.
set_weights(weights)	Sets the weights of the layer, from NumPy arrays.
summary()	Provide a summary of the learner's parameters.
with_name_scope(method)	Decorator to automatically enter the module name scope.

Attributes

activity_regularizer	Optional regularizer function for the output of this layer.
compute_dtype	The dtype of the layer's computations.
dtype	The dtype of the layer weights.
dtype_policy	The dtype policy associated with this layer.
dynamic	Whether the layer is dynamic (eager-only); set in the constructor.
inbound_nodes	Return Functional API nodes upstream of this layer.
input	Retrieves the input tensor(s) of a layer.
input_mask	Retrieves the input mask tensor(s) of a layer.
input_shape	Retrieves the input shape(s) of a layer.
input_spec	InputSpec instance(s) describing the input format for this layer.
losses	List of losses added using the add_loss() API.
metrics	List of metrics attached to the layer.
my_params
name	Name of the layer (string), set in the constructor.
name_scope	Returns a tf.name_scope instance for this class.
non_trainable_variables	Sequence of non-trainable variables owned by this module and its submodules.
non_trainable_weights	List of all non-trainable weights tracked by this layer.
outbound_nodes	Return Functional API nodes downstream of this layer.
output	Retrieves the output tensor(s) of a layer.
output_mask	Retrieves the output mask tensor(s) of a layer.
output_shape	Retrieves the output shape(s) of a layer.
stateful
submodules	Sequence of all sub-modules.
supports_masking	Whether this layer supports computing a mask using compute_mask.
trainable
trainable_variables	Sequence of trainable variables owned by this module and its submodules.
trainable_weights	List of all trainable weights tracked by this layer.
updates
variable_dtype	Alias of Layer.dtype, the dtype of the weights.
variables	Returns the list of all layer variables/weights.
weights	Returns the list of all layer variables/weights.

__init__(units, memory_size, num_heads)

Parameters:

• units (int)

• memory_size (int)

• num_heads (int)

build(input_shape)

Build method that creates the trainable memory matrix of shape (memory_size, units).

Parameters:

input_shape (tuple) – Shape of the input, e.g. (batch_size, time_steps, units).

call(inputs, training=False)

Forward pass of MemoryAugmentedAttention.

Parameters:

• inputs (tf.Tensor) – A 3D tensor of shape (batch_size, time_steps, units).

• training (bool, optional) – Indicates whether the layer is in training mode. Defaults to False.

Returns:

A tensor of the same shape as inputs: (batch_size, time_steps, units), augmented by the learned memory.

Return type:

tf.Tensor

get_config()

Returns configuration of this layer.

Returns:

Dictionary including ‘units’ and ‘memory_size’.

Return type:

dict

classmethod from_config(config)

Creates a new instance from a given config dictionary.

Parameters:

config (dict) – Configuration dictionary as returned by get_config.

Returns:

A new instance of this layer.

Return type:

MemoryAugmentedAttention

help(**kwargs)

my_params = MemoryAugmentedAttention(units, memory_size, num_heads)