Search papers, labs, and topics across Lattice.
IMT Atlantique
1
0
1
A new taxonomy reveals how distributional shifts in reinforcement learning can be systematically categorized, bridging the gap between ID/OOD generalization and non-stationary environments.