Search papers, labs, and topics across Lattice.
Affiliation:
3
0
5
Provably private tabular classification is now achievable with a single forward pass, outperforming baselines and drastically cutting fitting time.
Mapping pure differential privacy to Gaussian differential privacy reveals that a conservative $渭$ value can significantly enhance privacy without sacrificing performance.
Averaging membership inference scores across multiple individuals to reduce compute can lead to unreliable vulnerability assessments due to uncalibrated false positive rates.
