Mihail Plesa

Mihail Plesa

Teaching assistant at Military Technical Academy of Bucharest

Bio

Mihail is a teaching assistant at the Military Technical Academy of Bucharest where he teaches computer vision labs. His research area involves machine learning and cryptography. He is passion about problem solving, algorithms and data structures. Among other things, he publishes in the largest Medium publication about ML, Towards Data Science. 

A step by step approach on secure unsupervised machine learning using OPE

I think every ML engineer has worked with unsupervised learning algorithms like k-means or DBSCAN. But I don’t think most engineers have wondered what it would do if they had to upload customer data to the cloud in order to apply the algorithm. In this talk I will show you how Order Preserving Encryption (OPE) can be applied in order to guarantee the confidentiality of the data in the cloud without affecting the results of the clustering algorithm. Long story short: we encrypt the data, send data to the cloud, apply the clustering algorithm and finally download and decrypt the result. This presentation is a practical one and I intent to show you step by step which are the challenges encountered in implementing such a system and how they can be solved.