Google I/O is a wrap! Catch up on TensorFlow sessions View sessions

Module: tf_agents.bandits.agents

Module importing all agents.

Modules

bernoulli_thompson_sampling_agent module: An Thompson sampling agent for Bernoulli bandit problems.

dropout_thompson_sampling_agent module: A neural network based agent that implements Thompson sampling via dropout.

exp3_agent module: Implements the EXP3 bandit algorithm.

exp3_mixture_agent module: A mixture agent that updates the mixture distribution based on EXP3.

greedy_multi_objective_neural_agent module: An agent that uses and trains a greedy reward prediction policy.

greedy_reward_prediction_agent module: An agent that uses and trains a greedy reward prediction policy.

lin_ucb_agent module: Implements the Linear UCB bandit algorithm.

linear_bandit_agent module: An agent that maintains linear estimates for rewards and their uncertainty.

linear_thompson_sampling_agent module: Implements the Linear Thompson Sampling bandit algorithm.

mixture_agent module: An agent that mixes a list of agents with a constant mixture distribution.

neural_boltzmann_agent module: A neural network based agent that implements Boltzmann exploration.

neural_epsilon_greedy_agent module: A neural network based agent that implements epsilon greedy exploration.

neural_linucb_agent module: Implements the Neural + LinUCB bandit algorithm.

static_mixture_agent module: An agent that mixes a list of agents with a constant mixture distribution.

utils module: Common utility code and linear algebra functions.