Greedy policy Machine Learning
noun phrase
Definition: A policy that selects the action with the highest currently estimated value at each step (i.e., exploits current knowledge) [Google Machine Learning Glossary].
Example in context: “The maze has two non-colliding paths so that as soon as the agents split up, they can follow a locally greedy policy to get an optimal return.” [Lyu et al. 2023]
Related terms: epsilon-greedy, exploration, exploitation, policy