Private Text Classification
This addresses privacy concerns for organizations handling sensitive text data, but it is incremental as it builds on existing privacy techniques.
The paper tackled the problem of performing text classification on confidential corpora without sharing the data, by constructing text processing applications using privacy-preserving techniques like homomorphic encryption and secure computation, resulting in preliminary methods for binary classifiers.
Confidential text corpora exist in many forms, but do not allow arbitrary sharing. We explore how to use such private corpora using privacy preserving text analytics. We construct typical text processing applications using appropriate privacy preservation techniques (including homomorphic encryption, Rademacher operators and secure computation). We set out the preliminary materials from Rademacher operators for binary classifiers, and then construct basic text processing approaches to match those binary classifiers.