Search papers, labs, and topics across Lattice.
Affiliation:
1
0
2
Adversarial training gains a principled probabilistic interpretation and improved robustness via KL divergence maximization, outperforming standard methods.
