Differential Privacy in Natural Language Processing: The Story So Far

In an age where a vast amount of data is being Alas, in the field of NLP, where the core unit of produced daily, the opportunities created by this data is unstructured, fuzzy text rather than a structured proliferation increase concurrently. The availability data point, an initial attempt to apply Differential of big data enables countless downstream tasks Privacy poses some challenges. Chief among whose accuracy and utility seem to increase with these is the challenge of how to transfer the core the amount of data used. Specifically, the fields concepts of Differential Privacy, namely the "individual" of Machine Learning (ML) and Deep Learning and adjacency, to the textual domain where (DL) have profited from such data. Particularly in these concepts are not easily perceivable. Thus, it the case of Natural Language Processing (NLP), becomes the goal to find new ways of reasoning the tasks at hand more often than not concern the about Differential Privacy in order to adapt it to handling of unstructured data, meaning data that the unstructured data domain of NLP. Through the is not neatly organized into a traditional row-like course of this paper, the foundations of Differential database structure, and furthermore, data that is Privacy in the lens of NLP will be investigated, motivated not necessarily static. In fact, it is estimated that by some privacy vulnerabilities that surface data on the order of zettabytes (ZB) is being produced from NLP techniques. Afterwards, the limitations every day (Begum and Nausheen, 2018), and and open questions of Differential Privacy with within this amount, roughly 80% is unstructured, NLP will be analyzed with an in-depth discussion.