Decentralized Data Science

Unlocking Data Value, Respecting Privacy.

What you'll learn

Overview of Data Science and Machine Learning

Federated Learning

Decentralized Data Marketplaces

Differential Privacy

Homomorphic Encryption

TensorFlow Federated (TFF)

TensorFlow Lite

Requirements

Some basic understanding of data science and machine learning is required to take this course.

Description

Please note that this is not a Data Science or Machine Learning course. This course does not cover any coding. Welcome to the course on "Decentralized Data Science" – an exploration into the intersection of cutting-edge technologies and the transformative power of decentralized approaches in Data Science - especially in Machine Learning. ChatGPT brought us to the verge of an AI Race. It is expected that in the coming months and years, all the tech majors will launch many new AI models. We are all excited about the sector that is poised for dramatic innovation. But, is there anything we should be concerned about? Yes. Privacy.These tech majors are likely to use user data to train their models. As centralized data processing involves various vulnerabilities, user privacy will be at stake in this AI Race. So, is there any way to preserve user privacy in Machine Learning? This is where Decentralized Data Science comes in. Decentralized Machine Learning offers various frameworks such as Federated Learning, Differential Privacy, Homomorphic Encryption, Secure Multi-Party Computations, and Edge Computing. These frameworks enable processing of data while preserving user privacy. We will also discuss tools such as TensorFlow Federated and TensorFlow Lite that help us build these decentralized machine learning systems. Let us discuss these concepts in this course

Overview

Section 1: Introduction

Lecture 1 Introduction

Lecture 2 Who is this course for?

Lecture 3 Course Outline

Section 2: Basics of Data Science

Lecture 4 What is Data Science?

Lecture 5 Classification of Data Science

Section 3: Primer on Machine Learning

Lecture 6 Introduction

Lecture 7 Machine Learning Models

Lecture 8 Representation of ML Models

Lecture 9 ML Training

Lecture 10 ML Frameworks

Section 4: MLOps

Lecture 11 Introduction

Lecture 12 Overview of MLOps

Section 5: Why does data science need to be decentralized?

Lecture 13 Why does data science need to be decentralized?

Section 6: Federated Learning

Lecture 14 Introduction

Lecture 15 TensorFlow Federated (TFF)

Lecture 16 Federated Averaging (FedAvg)

Lecture 17 Secure Aggregation

Lecture 18 TensorFlow Lite

Lecture 19 Federated Datasets

Lecture 20 Federated optimization

Lecture 21 Use Cases

Section 7: Decentralized Data Marketplaces

Lecture 22 Introduction

Lecture 23 Workings

Section 8: Differential Privacy

Lecture 24 Differential Privacy

Section 9: Homomorphic Encryption

Lecture 25 Introduction

Lecture 26 Use Cases

Section 10: Edge Computing and Edge Analytics

Lecture 27 Introduction

Lecture 28 Federated Learning Vs Edge Analytics

Lecture 29 Edge Analytics Use Cases

Lecture 30 Use of Edge Computing with Federated Learning

Section 11: Secure Multi-Party Computation (SMPC)

Lecture 31 Introduction

Lecture 32 Protocols

Section 12: Tensorflow Federated (TFF)

Lecture 33 Introduction

Lecture 34 TensorFlow Federated APIs

Lecture 35 Example Application - Federated Learning (FL) API

Lecture 36 Example Application - Federated Core (FC) API

Section 13: TensorFlow Lite

Lecture 37 Introduction

Lecture 38 Role in Decentralized Data Science

Lecture 39 Sample Application

Section 14: Thank You

Lecture 40 Thank You

Techies and Tech Investors