Search papers, labs, and topics across Lattice.
1
0
Ditching pessimism unlocks a quadratically faster, $\widetilde{\mathcal{O}}(1/n)$ sample complexity for offline learning in KL-regularized games.
