"""A collection of variables used by `NeuralLinUCBAgent`."""

def __init__(

self,

num_actions: int,

encoding_dim: int,

dtype: tf.DType = tf.float64,

name: Optional[Text] = None,

):

"""Initializes an instance of `NeuralLinUCBVariableCollection`.

Args:

num_actions: (int) number of actions the agent acts on.

encoding_dim: (int) The dimensionality of the output of the encoding

network.

dtype: The type of the variables. Should be one of `tf.float32` and

`tf.float64`.

name: (string) the name of this instance.

"""

tf.Module.__init__(self, name=name)

self.actions_from_reward_layer = tf.compat.v2.Variable(

True, dtype=tf.bool, name='is_action_from_reward_layer'

)

self.cov_matrix_list = []

self.data_vector_list = []

# We keep track of the number of samples per arm.

self.num_samples_list = []

for k in range(num_actions):

self.cov_matrix_list.append(

tf.compat.v2.Variable(

tf.zeros([encoding_dim, encoding_dim], dtype=dtype),

name='a_{}'.format(k),

)

)

self.data_vector_list.append(

tf.compat.v2.Variable(

tf.zeros(encoding_dim, dtype=dtype), name='b_{}'.format(k)

)

)

self.num_samples_list.append(

tf.compat.v2.Variable(

tf.zeros([], dtype=dtype), name='num_samples_{}'.format(k)

)

"""An agent implementing the LinUCB algorithm on top of a neural network."""

def __init__(

self,

time_step_spec: types.TimeStep,

action_spec: types.BoundedTensorSpec,

encoding_network: types.Network,

encoding_network_num_train_steps: int,

encoding_dim: int,

optimizer: types.Optimizer,

variable_collection: Optional[NeuralLinUCBVariableCollection] = None,

alpha: float = 1.0,

gamma: float = 1.0,

epsilon_greedy: float = 0.0,

observation_and_action_constraint_splitter: Optional[

types.Splitter

] = None,

accepts_per_arm_features: bool = False,

distributed_train_encoding_network: bool = False,

# Params for training.

error_loss_fn: types.LossFn = tf.compat.v1.losses.mean_squared_error,

gradient_clipping: Optional[float] = None,

# Params for debugging.

debug_summaries: bool = False,

summarize_grads_and_vars: bool = False,

train_step_counter: Optional[tf.Variable] = None,

emit_policy_info: Sequence[Text] = (),

emit_log_probability: bool = False,

dtype: tf.DType = tf.float64,

name: Optional[Text] = 'neural_linucb_agent',

):

"""Initialize an instance of `NeuralLinUCBAgent`.

Args:

time_step_spec: A `TimeStep` spec describing the expected `TimeStep`s.

action_spec: A scalar `BoundedTensorSpec` with `int32` or `int64` dtype

describing the number of actions for this agent.

encoding_network: a Keras network that encodes the observations.

encoding_network_num_train_steps: how many training steps to run for

training the encoding network before switching to LinUCB. If negative,

the encoding network is assumed to be already trained.

encoding_dim: the dimension of encoded observations.

optimizer: The optimizer to use for training.

variable_collection: Instance of `NeuralLinUCBVariableCollection`.

Collection of variables to be updated by the agent. If `None`, a new

instance of `LinearBanditVariables` will be created. Note that this

collection excludes the variables owned by the encoding network.

alpha: (float) positive scalar. This is the exploration parameter that

multiplies the confidence intervals.

gamma: a float forgetting factor in [0.0, 1.0]. When set to 1.0, the

algorithm does not forget.

epsilon_greedy: A float representing the probability of choosing a random

action instead of the greedy action.

observation_and_action_constraint_splitter: A function used for masking

valid/invalid actions with each state of the environment. The function

takes in a full observation and returns a tuple consisting of 1) the

part of the observation intended as input to the bandit agent and

policy, and 2) the boolean mask. This function should also work with a

`TensorSpec` as input, and should output `TensorSpec` objects for the

observation and mask.

accepts_per_arm_features: (bool) Whether the policy accepts per-arm

features.

distributed_train_encoding_network: (bool) whether to train the encoding

network or not. This applies only in distributed training setting. When

set to true this agent will train the encoding network. Otherwise, it

will assume the encoding network is already trained and will train

LinUCB on top of it.

error_loss_fn: A function for computing the error loss, taking parameters

labels, predictions, and weights (any function from tf.losses would

work). The default is `tf.losses.mean_squared_error`.

gradient_clipping: A float representing the norm length to clip gradients

(or None for no clipping.)

debug_summaries: A Python bool, default False. When True, debug summaries

are gathered.

summarize_grads_and_vars: A Python bool, default False. When True,

gradients and network variable summaries are written during training.

train_step_counter: An optional `tf.Variable` to increment every time the

train op is run. Defaults to the `global_step`.

emit_policy_info: (tuple of strings) what side information we want to get

as part of the policy info. Allowed values can be found in

`policy_utilities.PolicyInfo`.

emit_log_probability: Whether the NeuralLinUCBPolicy emits

log-probabilities or not. Since the policy is deterministic, the

probability is just 1.

dtype: The type of the parameters stored and updated by the agent. Should

be one of `tf.float32` and `tf.float64`. Defaults to `tf.float64`.

name: a name for this instance of `NeuralLinUCBAgent`.

Raises:

TypeError if variable_collection is not an instance of

`NeuralLinUCBVariableCollection`.

ValueError if dtype is not one of `tf.float32` or `tf.float64`.

"""

tf.Module.__init__(self, name=name)

common.tf_agents_gauge.get_cell('TFABandit').set(True)

self._num_actions = policy_utilities.get_num_actions_from_tensor_spec(

action_spec

)

self._num_models = 1 if accepts_per_arm_features else self._num_actions

self._observation_and_action_constraint_splitter = (

observation_and_action_constraint_splitter

)

self._accepts_per_arm_features = accepts_per_arm_features

self._alpha = alpha

if variable_collection is None:

variable_collection = NeuralLinUCBVariableCollection(

self._num_models, encoding_dim, dtype

)

elif not isinstance(variable_collection, NeuralLinUCBVariableCollection):

raise TypeError(

'Parameter `variable_collection` should be '

'of type `NeuralLinUCBVariableCollection`.'

)

self._variable_collection = variable_collection

self._gamma = gamma

if self._gamma < 0.0 or self._gamma > 1.0:

raise ValueError('Forgetting factor `gamma` must be in [0.0, 1.0].')

self._dtype = dtype

if dtype not in (tf.float32, tf.float64):

raise ValueError(

'Agent dtype should be either `tf.float32 or `tf.float64`.'

)

self._epsilon_greedy = epsilon_greedy

reward_layer = tf.keras.layers.Dense(

self._num_models,

kernel_initializer=tf.random_uniform_initializer(

minval=-0.03, maxval=0.03

),

use_bias=False,

activation=None,

name='reward_layer',

)

encoding_network.create_variables()

self._encoding_network = encoding_network

reward_layer.build(input_shape=tf.TensorShape([None, encoding_dim]))

self._reward_layer = reward_layer

self._encoding_network_num_train_steps = encoding_network_num_train_steps

self._encoding_dim = encoding_dim

self._optimizer = optimizer

self._error_loss_fn = error_loss_fn

self._gradient_clipping = gradient_clipping

train_step_counter = tf.compat.v1.train.get_or_create_global_step()

self._distributed_train_encoding_network = (

distributed_train_encoding_network

)

policy = neural_linucb_policy.NeuralLinUCBPolicy(

encoding_network=self._encoding_network,

encoding_dim=self._encoding_dim,

reward_layer=self._reward_layer,

epsilon_greedy=self._epsilon_greedy,

actions_from_reward_layer=self.actions_from_reward_layer,

cov_matrix=self.cov_matrix,

data_vector=self.data_vector,

num_samples=self.num_samples,

time_step_spec=time_step_spec,

alpha=alpha,

emit_policy_info=emit_policy_info,

emit_log_probability=emit_log_probability,

accepts_per_arm_features=accepts_per_arm_features,

distributed_use_reward_layer=distributed_train_encoding_network,

observation_and_action_constraint_splitter=(

observation_and_action_constraint_splitter

),

)

training_data_spec = None

if accepts_per_arm_features:

training_data_spec = bandit_spec_utils.drop_arm_observation(

policy.trajectory_spec

)

super(NeuralLinUCBAgent, self).__init__(

time_step_spec=time_step_spec,

action_spec=policy.action_spec,

policy=policy,

collect_policy=policy,

train_sequence_length=None,

training_data_spec=training_data_spec,

debug_summaries=debug_summaries,

summarize_grads_and_vars=summarize_grads_and_vars,

train_step_counter=train_step_counter,

)

self._as_trajectory = data_converter.AsTrajectory(

self.data_context, sequence_length=None

)

@property

def num_actions(self):

return self._num_actions

@property

def actions_from_reward_layer(self):

return self._variable_collection.actions_from_reward_layer

@property

def cov_matrix(self):

return self._variable_collection.cov_matrix_list

@property

def data_vector(self):

return self._variable_collection.data_vector_list

@property

def num_samples(self):

return self._variable_collection.num_samples_list

@property

def alpha(self):

return self._alpha

@property

def variables(self):

return (

self.num_samples

+ self.cov_matrix

+ self.data_vector

+ self._encoding_network.trainable_weights

+ self._reward_layer.trainable_weights

+ [self.train_step_counter]

)

@alpha.setter

def update_alpha(self, alpha):

return tf.compat.v1.assign(self._alpha, alpha)

def _initialize(self):

tf.compat.v1.variables_initializer(self.variables)

def compute_summaries(self, loss):

with tf.name_scope('Losses/'):

tf.compat.v2.summary.scalar(

name='total_loss', data=loss, step=self.train_step_counter

)

if self._summarize_grads_and_vars:

with tf.name_scope('Variables/'):

trainable_variables = (

self._encoding_network.trainable_weights

+ self._reward_layer.trainable_weights

)

for var in trainable_variables:

tf.compat.v2.summary.histogram(

name=var.name.replace(':', '_'),

data=var,

step=self.train_step_counter,

)

def _loss_using_reward_layer(

self,

observations: types.NestedTensor,

actions: types.Tensor,

rewards: types.Tensor,

weights: Optional[types.Float] = None,

training: bool = False,

) -> tf_agent.LossInfo:

"""Computes loss for reward prediction training.

Args:

observations: A batch of observations.

actions: A batch of actions.

rewards: A batch of rewards.

weights: Optional scalar or elementwise (per-batch-entry) importance

weights. The output batch loss will be scaled by these weights, and the

final scalar loss is the mean of these values.

training: Whether the loss is being used for training.

Returns:

loss: A `LossInfo` containing the loss for the training step.

"""

with tf.name_scope('loss'):

encoded_observation, _ = self._encoding_network(

observations, training=training

)

encoded_observation = tf.reshape(

encoded_observation, shape=[-1, self._encoding_dim]

)

predicted_rewards = self._reward_layer(

encoded_observation, training=training

)

chosen_actions_predicted_rewards = common.index_with_actions(

predicted_rewards, tf.cast(actions, dtype=tf.int32)

)

loss = self._error_loss_fn(

rewards,

chosen_actions_predicted_rewards,

1 if weights is None else weights,

)

if self._summarize_grads_and_vars:

with tf.name_scope('Per_arm_loss/'):

for k in range(self._num_models):

loss_mask_for_arm = tf.cast(tf.equal(actions, k), tf.float32)

loss_for_arm = self._error_loss_fn(

rewards,

chosen_actions_predicted_rewards,

weights=loss_mask_for_arm,

)

tf.compat.v2.summary.scalar(

name='loss_arm_' + str(k),

data=loss_for_arm,

step=self.train_step_counter,

)

return tf_agent.LossInfo(loss, extra=())

def compute_loss_using_reward_layer(

self,

observation: types.NestedTensor,

action: types.Tensor,

reward: types.Tensor,

weights: Optional[types.Float] = None,

training: bool = False,

) -> tf_agent.LossInfo:

"""Computes loss using the reward layer.

Args:

observation: A batch of observations.

action: A batch of actions.

reward: A batch of rewards.

weights: Optional scalar or elementwise (per-batch-entry) importance

weights. The output batch loss will be scaled by these weights, and the

final scalar loss is the mean of these values.

training: Whether the loss is being used for training.

Returns:

loss: A `LossInfo` containing the loss for the training step.

"""

# Update the neural network params.

with tf.GradientTape() as tape:

loss_info = self._loss_using_reward_layer(

observation, action, reward, weights, training=training

)

tf.debugging.check_numerics(loss_info[0], 'Loss is inf or nan')

tf.compat.v2.summary.scalar(

name='using_reward_layer', data=1, step=self.train_step_counter

)

if self._summarize_grads_and_vars:

self.compute_summaries(loss_info.loss)

variables_to_train = (

self._encoding_network.trainable_weights

+ self._reward_layer.trainable_weights

)

if not variables_to_train:

raise ValueError('No variable to train in the agent.')

grads = tape.gradient(loss_info.loss, variables_to_train)

grads_and_vars = tuple(zip(grads, variables_to_train))

if self._gradient_clipping is not None:

grads_and_vars = eager_utils.clip_gradient_norms(

grads_and_vars, self._gradient_clipping

)

if self._summarize_grads_and_vars:

with tf.name_scope('Reward_network/'):

eager_utils.add_variables_summaries(

grads_and_vars, self.train_step_counter

)

eager_utils.add_gradients_summaries(

grads_and_vars, self.train_step_counter

)

self._optimizer.apply_gradients(grads_and_vars)

self.train_step_counter.assign_add(1)

return loss_info

def compute_loss_using_linucb(

self,

observation: types.NestedTensor,

action: types.Tensor,

reward: types.Tensor,

weights: Optional[types.Float] = None,

training: bool = False,

) -> tf_agent.LossInfo:

"""Computes the loss using LinUCB.

Args:

observation: A batch of observations.

action: A batch of actions.

reward: A batch of rewards.

weights: unused weights.

training: Whether the loss is being used to train.

Returns:

loss: A `LossInfo` containing the loss for the training step.

"""

del weights # unused

# The network is trained now. Update the covariance matrix.

encoded_observation, _ = self._encoding_network(

observation, training=training

)

encoded_observation = tf.cast(encoded_observation, dtype=self._dtype)

encoded_observation = tf.reshape(

encoded_observation, shape=[-1, self._encoding_dim]

)

for k in range(self._num_models):

diag_mask = tf.linalg.tensor_diag(

tf.cast(tf.equal(action, k), self._dtype)

)

observations_for_arm = tf.matmul(diag_mask, encoded_observation)

rewards_for_arm = tf.matmul(diag_mask, tf.reshape(reward, [-1, 1]))

num_samples_for_arm_current = tf.reduce_sum(diag_mask)

tf.compat.v1.assign_add(self.num_samples[k], num_samples_for_arm_current)

num_samples_for_arm_total = self.num_samples[k].read_value()

# Update the matrix A and b.

# pylint: disable=cell-var-from-loop

def update(cov_matrix, data_vector):

a_new, b_new = linear_agent.update_a_and_b_with_forgetting(

cov_matrix,

data_vector,

rewards_for_arm,

observations_for_arm,

self._gamma,

)

return a_new, b_new

a_new, b_new = tf.cond(

tf.squeeze(num_samples_for_arm_total) > 0,

lambda: update(self.cov_matrix[k], self.data_vector[k]),

lambda: (self.cov_matrix[k], self.data_vector[k]),

)

tf.compat.v1.assign(self.cov_matrix[k], a_new)

tf.compat.v1.assign(self.data_vector[k], b_new)

loss_tensor = tf.cast(-1.0 * tf.reduce_sum(reward), dtype=tf.float32)

loss_info = tf_agent.LossInfo(loss=loss_tensor, extra=())

tf.compat.v2.summary.scalar(

name='using_reward_layer', data=0, step=self.train_step_counter

)

self.train_step_counter.assign_add(1)

return loss_info

def compute_loss_using_linucb_distributed(

self,

observation: types.NestedTensor,

action: types.Tensor,

reward: types.Tensor,

weights: Optional[types.Float] = None,

training: bool = False,

) -> tf_agent.LossInfo:

"""Computes the loss using LinUCB distributively.

Args:

observation: A batch of observations.

action: A batch of actions.

reward: A batch of rewards.

weights: unused weights.

training: Whether the loss is being used to train.

Returns:

loss: A `LossInfo` containing the loss for the training step.

"""

del weights # unused

# The network is trained now. Update the covariance matrix.

encoded_observation, _ = self._encoding_network(

observation, training=training

)

encoded_observation = tf.cast(encoded_observation, dtype=self._dtype)

encoded_observation = tf.reshape(

encoded_observation, shape=[-1, self._encoding_dim]

)

self._train_step_counter.assign_add(1)

for k in range(self._num_models):

diag_mask = tf.linalg.tensor_diag(

tf.cast(tf.equal(action, k), self._dtype)

)

observations_for_arm = tf.matmul(diag_mask, encoded_observation)

rewards_for_arm = tf.matmul(diag_mask, tf.reshape(reward, [-1, 1]))

# Compute local updates for the matrix A and b of this arm.

cov_matrix_local_udpate = tf.matmul(

observations_for_arm, observations_for_arm, transpose_a=True

)

data_vector_local_update = bandit_utils.sum_reward_weighted_observations(

rewards_for_arm, observations_for_arm

)

def _merge_fn(

strategy,

per_replica_cov_matrix_update,

per_replica_data_vector_update,

):

"""Merge the per-replica-updates."""

# Reduce the per-replica-updates using SUM.

# pylint: disable=cell-var-from-loop

updates_and_vars = [

(per_replica_cov_matrix_update, self.cov_matrix[k]),

(per_replica_data_vector_update, self.data_vector[k]),

]

reduced_updates = strategy.extended.batch_reduce_to(

tf.distribute.ReduceOp.SUM, updates_and_vars

)

# Update the model variables.

self.cov_matrix[k].assign(

self._gamma * self.cov_matrix[k] + reduced_updates[0]

)

self.data_vector[k].assign(

self._gamma * self.data_vector[k] + reduced_updates[1]

)

# Passes the local_updates to the _merge_fn() above that performs custom

# computation on the per-replica values.

# All replicas pause their execution until merge_call() is done and then,

# execution is resumed.

replica_context = tf.distribute.get_replica_context()

replica_context.merge_call(

_merge_fn, args=(cov_matrix_local_udpate, data_vector_local_update)

)

loss = -1.0 * tf.reduce_sum(reward)

return tf_agent.LossInfo(loss=(loss), extra=())

def _train(self, experience, weights=None):

"""Updates the policy based on the data in `experience`.

Note that `experience` should only contain data points that this agent has

not previously seen. If `experience` comes from a replay buffer, this buffer

should be cleared between each call to `train`.

Args:

experience: A batch of experience data in the form of a `Trajectory`.

weights: (optional) sample weights.

Returns:

A `LossInfo` containing the loss *before* the training step is taken.

In most cases, if `weights` is provided, the entries of this tuple will

have been calculated with the weights. Note that each Agent chooses

its own method of applying weights.

"""

experience = self._as_trajectory(experience)

(observation, action, reward) = (

bandit_utils.process_experience_for_neural_agents(

experience, self._accepts_per_arm_features, self.training_data_spec

)

)

if self._observation_and_action_constraint_splitter is not None:

observation, _ = self._observation_and_action_constraint_splitter(

observation

)

reward = tf.cast(reward, self._dtype)

if tf.distribute.has_strategy():

if self._distributed_train_encoding_network:

loss_info = self.compute_loss_using_reward_layer(

observation, action, reward, weights, training=True

)

else:

loss_info = self.compute_loss_using_linucb_distributed(

observation, action, reward, weights, training=True

)

return loss_info

tf.compat.v1.assign(

self.actions_from_reward_layer,

tf.less(

self._train_step_counter, self._encoding_network_num_train_steps

),

)

def use_actions_from_reward_layer():

return self.compute_loss_using_reward_layer(

observation, action, reward, weights, training=True

)

def no_actions_from_reward_layer():

return self.compute_loss_using_linucb(

observation, action, reward, weights, training=True

)

loss_info = tf.cond(

self.actions_from_reward_layer,

use_actions_from_reward_layer,

no_actions_from_reward_layer,

)