Search papers, labs, and topics across Lattice.
RWTH Aachen University
2
0
RL agents can learn to stay safe in complex environments without excessive conservatism by explicitly modeling and avoiding regions of high uncertainty.