Search papers, labs, and topics across Lattice.
1
0
3
Achieve state-of-the-art RL performance on DeepMind Control Suite by dynamically adjusting relative entropy regularization and using multi-scale state-action embeddings.
