Decentralized Data Science

Unlocking Data Value, Respecting Privacy.

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