Tabular Q-learning Machine Learning
noun phrase
Definition: A form of Q-learning in which the action-value function is represented explicitly as a table of Q-values for each state–action pair, rather than approximated by a parametric model such as a neural network [Singi et al. 2023].
Example in context: “This estimator can be of the same type as used for Q: for tabular Q-learning, we can use a table, while for Deep Q-Networks (DQN) we use a deep neural network.” [Singi et al. 2023]
Related terms: Q-learning; reinforcement learning; Q-table; Deep Q-Network (DQN); function approximation