©SHUTTERSTOCK.COM/THAPANA_STUDIO Privacy-Preserving Deep Learning With Homomorphic Encryption: An Introduction Alessandro Falcetta and Manuel Roveri Politecnico di Milano, ITALY Abstract-Privacy-preserving deep learning with homomorphic encryption (HE) is a novel and promising research area aimed at designing deep learning solutions that operate while guaranteeing the privacy of user data. Designing privacy-preserving deep learning solutions requires one to completely rethink and redesign deep learning models and algorithms to match the severe technological and algorithmic constraints of HE. This paper provides an introduction to this complex research area as well as a Digital Object Identifier 10.1109/MCI.2022.3180883 Date of current version: 19 July 2022 methodology for designing privacy-preserving convolutional neural networks (CNNs). This methodology was applied to the design of a privacy-preserving version of the well-known LeNet-1 CNN, which was successfully operated on two benchmark datasets for image classification. Furthermore, this paper details and comments on the research challenges and software resources available for privacy-preserving deep learning with HE. I. Introduction T 14 IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE | AUGUST 2022 oday's world is characterized by information abundance [1]. Thousands of exabytes of data are generated every day [2] by Internet-of-Things systems, mobile devices, social media, and industrial machinery. To extract value from these data, intelligent " data-processing " services have increased in number in recent years, which are based Corresponding author: Alessandro Falcetta (e-mail: alessandro.falcetta@polimi.it). 1556-603X/22©2022IEEEhttp://www.SHUTTERSTOCK.COM/THAPANA_STUDIO