Sheng Zha

Automatic Clipping: Differentially Private Deep Learning Made Easier and Stronger

2022 · arXiv (Cornell University)

Per-example gradient clipping is a key algorithmic step that enables practical differential private (DP) training for deep learning models. The choice of clipping threshold R, however, is vital for achieving high accuracy under DP. We …

Unlearn Dataset Bias in Natural Language Inference by Fitting the Residual

2019

Statistical natural language inference (NLI) models are susceptible to learning dataset bias: superficial cues that happen to associate with the label on a particular dataset, but are not useful in general, e.g., negation words indicate …