DP-NMT: Scalable Differentially-Private Machine Translation

arXiv:2311.14465v2 Announce Type: replace
Abstract: Neural machine translation (NMT) is a widely popular text generation task, yet there is a considerable research gap in the development of privacy-preserving NMT models, despite significant data privacy concerns for NMT systems. Differentially private stochastic gradient descent (DP-SGD) is a popular method for training machine learning models with concrete privacy guarantees; however, the implementation specifics of training a model with DP-SGD are not always clarified in existing models, with differing software libraries used and …

abstract arxiv concerns concrete cs.cl data data privacy development gap gradient machine machine learning machine learning models machine translation neural machine translation popular privacy research scalable stochastic systems text text generation training translation type