Decision Transformer

Trajectory Modeling

Approach involving modeling complete trajectories (states, actions, rewards) as continuous sequences for policy learning in offline RL.

← Indietro