Learning and Evaluating a Differentially Private Pre-trained Language Model

Shlomo Hoory, Amir Feder, Avichai Tendler, Alon Cohen, Sofia Erell, Itay Laish, Hootan Nakhost, Uri Stemmer, Ayelet Benjamini, Avinatan Hassidim, Yossi Matias

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Contextual language models have led to significantly better results, especially when pretrained on the same data as the downstream task. While this additional pre-training usually improves performance, it can lead to information leakage and therefore risks the privacy of individuals mentioned in the training data. One method to guarantee the privacy of such individuals is to train a differentially-private language model, but this usually comes at the expense of model performance. Also, in the absence of a differentially private vocabulary training, it is not possible to modify the vocabulary to fit the new data, which might further degrade results. In this work we bridge these gaps, and provide guidance to future researchers and practitioners on how to improve privacy while maintaining good model performance. We introduce a novel differentially private word-piece algorithm, which allows training a tailored domain-specific vocabulary while maintaining privacy. We then experiment with entity extraction tasks from clinical notes, and demonstrate how to train a differentially private pre-trained language model (i.e., BERT) with a privacy guarantee of ? = 1.1 and with only a small degradation in performance. Finally, as it is hard to tell given a privacy parameter ? what was the effect on the trained representation, we present experiments showing that the trained model does not memorize private information.

Original languageEnglish
Title of host publicationFindings of the Association for Computational Linguistics, Findings of ACL
Subtitle of host publicationEMNLP 2021
EditorsMarie-Francine Moens, Xuanjing Huang, Lucia Specia, Scott Wen-Tau Yih
PublisherAssociation for Computational Linguistics (ACL)
Pages1178-1189
Number of pages12
ISBN (Electronic)9781955917100
StatePublished - 2021
Externally publishedYes
Event2021 Findings of the Association for Computational Linguistics, Findings of ACL: EMNLP 2021 - Punta Cana, Dominican Republic
Duration: 7 Nov 202111 Nov 2021

Publication series

NameFindings of the Association for Computational Linguistics, Findings of ACL: EMNLP 2021

Conference

Conference2021 Findings of the Association for Computational Linguistics, Findings of ACL: EMNLP 2021
Country/TerritoryDominican Republic
CityPunta Cana
Period7/11/2111/11/21

Fingerprint

Dive into the research topics of 'Learning and Evaluating a Differentially Private Pre-trained Language Model'. Together they form a unique fingerprint.

Cite this