Towards Quantifying The Privacy Of Redacted Text

TL;DR

This paper introduces a method to quantify the privacy of redacted text by using deep learning to generate and analyze possible original texts, assessing privacy based on the diversity and similarity of reconstructed outputs.

Contribution

It presents a novel approach combining transformer-based models and k-anonymity concepts to evaluate privacy in redacted text, which was not previously addressed.

Findings

Effective reconstruction of original text from redacted versions.

Quantitative measure of privacy based on diversity of reconstructions.

Demonstrates the method's ability to assess privacy levels in redacted documents.

Abstract

In this paper we propose use of a k-anonymity-like approach for evaluating the privacy of redacted text. Given a piece of redacted text we use a state of the art transformer-based deep learning network to reconstruct the original text. This generates multiple full texts that are consistent with the redacted text, i.e. which are grammatical, have the same non-redacted words etc, and represents each of these using an embedding vector that captures sentence similarity. In this way we can estimate the number, diversity and quality of full text consistent with the redacted text and so evaluate privacy.