Federated Learning with PySyft

Jidin Dinesh (~jidin)


What if I tell you, we were being robbed off the most valuable natural resource of the 21st century, Data?

The future of Machine Learning is Decentralized.

Privacy and ML are not foes (against each other), even though the current AI business model makes it seem so. What if they can be allies? (No, not in a parallel universe. Here in this realm of ML)

Layout of talk

  • How the present AI business model is flawed (10 mins)
    • Cons of Centralized ML
    • Win-Win way of doing ML
  • Future of Privacy Preserving ML (10 mins)
    • Intro to Federated Learning
    • Intro to Differential Privacy
  • Code demo in PySyft (5 mins)


  1. Basics of ML , just being a ML enthusiast would suffice.

  2. Ability to read and make sense of PyTorch code.

  3. Laptop with Internet connectivity to run the shared Colab notebooks (optional)

Speaker Info:

Jidin Dinesh is a member of OpenMined.

Senior CSE undergrad at Cochin University of Science And Technology.

An aspiring AI researcher at 🧠 and a passionate photographer at ❤️.

Id: 1230
Section: Data Science, Machine Learning and AI
Type: Talks
Target Audience: Intermediate
Last Updated: