Online Optimization

Learning with Partial Information

Paradigm where the algorithm only receives information about the chosen action (bandit) rather than all possible actions (full information).

← Wstecz