Название: Homomorphic Encryption for Data Science (HE4DS)Автор: Allon Adir, Ehud Aharoni, Nir Drucker, Ronen Levy
Издательство: Springer
Год: 2024
Страниц: 311
Язык: английский
Формат: pdf (true), epub
Размер: 34.0 MB
This book provides basic knowledge required by an application developer to understand and use the Fully Homomorphic Encryption (FHE) technology for privacy preserving Data-Science applications. The authors present various techniques to leverage the unique features of FHE and to overcome its characteristic limitations. Homomorphic encryption (HE) is a cryptographic primitive that provides unique security guarantees in the privacy enhancing technologies (PETs) ecosystem. HE is a special type of encryption that allows computations to be performed on encrypted data. For example, it enables additions, multiplications, or both on ciphertexts, where the resulting ciphertext can be decrypted to have the same value as if the mathematical operations were performed directly on the encrypted data. This property is called homomorphism, hence the name of the HE primitive.
Specifically, this book summarizes polynomial approximation techniques used by FHE applications and various data packing schemes based on a data structure called tile tensors, and demonstrates how to use the studied techniques in several specific privacy preserving applications. Examples and exercises are also included throughout this book.
The proliferation of practical FHE technology has triggered a wide interest in the field and a common wish to experience and understand it. This book aims to simplify the FHE world for those who are interested in privacy preserving Data Science tasks, and for an audience that does not necessarily have a deep cryptographic background, including undergraduate and graduate-level students in Computer Science, and data scientists who plan to work on private data and models.
Data Science is the study of extrapolating insights out of data and information. It leverages tools and techniques from different academic fields such as mathematics, Computer Science, information science, and domain knowledge to analyze data and create data-driven observations, hypotheses, and conclusions. These hypotheses, or models, attempt to represent the underlying laws that govern the patterns we see. Using Data Science, we can now attempt to understand the data at a deeper level and even attempt to identify root causes that created the observed and collected data. Training and continuously updating the data models with different observations fine-tunes these models and can help to better approximate our understanding of the recorded phenomena. Thus, models can consequently be used to classify, infer, and predict new data points that were not part of the original training data. Since data is often representative of sensitive, private, or confidential information, its accumulation, processing, and insights are sometimes private as well. As such, data scientists are usually asked to adhere to various privacy regulations that restrict and regulate Data Science methodologies. The Chapter 1 provides a basic introduction to Data Science, data privacy concerns, and an overview of privacy-preserving techniques that attempt to address these concerns.
Скачать Homomorphic Encryption for Data Science (HE4DS)