Search papers, labs, and topics across Lattice.
1
0
2
0
MBRL agents can dramatically improve sample efficiency by using their world models not just for training, but also for actively planning to explore informative, high-entropy states at inference time.
