Search papers, labs, and topics across Lattice.
1
0
3
Overcome the prohibitive cost of ground-truth labels in reinforcement learning by actively acquiring labels for only the most valuable samples, leading to stable training and improved performance even with limited annotation budgets.
