Deployment of Differential Privacy for Application in Artificial Intelligence

Artificial Intelligence (AI) enables computers to mimic human intelligence. Some of the sub-fields of AI include Machine Learning (ML), Computer Vision (CV), and Natural Language Processing (NLP). In essence, ML enables computers to learn from data. On the other hand, CV mimics human vision by enabling computers to derive information from visual inputs such as images and videos. Finally, NLP makes it possible for computers to process the natural language used by humans. Some applications of AI sub-fields such as ML, CV, and NLP might involve the use of sensitive datasets; in which case the need for privacy is pronounced. One of the privacy tools that can be used to address the privacy concern in AI is Differential Privacy (DP); a privacy framework with mathematical guarantees of data privacy. This paper reports the deployment of Differential Privacy for application in privacy-preserving Machine Learning. Furthermore, this paper demonstrated the use of DP with various ML models. The results obtained from the work reported in this paper demonstrate the utility of using DP in AI; in order to address the data privacy concern.