Generalised Differential Privacy for Text Document Processing
This addresses privacy concerns for text data users, such as authors or organizations, by providing a method to anonymize writing style without losing content utility, though it is incremental as it builds on existing differential privacy and text processing ideas.
The paper tackles the problem of obfuscating stylistic clues in text documents to protect authorship while preserving content, by combining generalized differential privacy with machine learning techniques. It demonstrates on a fan fiction dataset that the method effectively disguises writing style while maintaining accuracy for content classification tasks.
We address the problem of how to "obfuscate" texts by removing stylistic clues which can identify authorship, whilst preserving (as much as possible) the content of the text. In this paper we combine ideas from "generalised differential privacy" and machine learning techniques for text processing to model privacy for text documents. We define a privacy mechanism that operates at the level of text documents represented as "bags-of-words" - these representations are typical in machine learning and contain sufficient information to carry out many kinds of classification tasks including topic identification and authorship attribution (of the original documents). We show that our mechanism satisfies privacy with respect to a metric for semantic similarity, thereby providing a balance between utility, defined by the semantic content of texts, with the obfuscation of stylistic clues. We demonstrate our implementation on a "fan fiction" dataset, confirming that it is indeed possible to disguise writing style effectively whilst preserving enough information and variation for accurate content classification tasks.