Federated ML on iOS: Protecting User Data in the Age of AI
summary
Federated learning on iOS: using CoreML and coremltools to train models locally on users’ devices instead of centrally, reducing data-leak risk. Covers federated vs centralized architectures, practical integration of a model into an iOS app for network-traffic analysis, plus delivering model updates and addressing security and compute-optimization challenges.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:05Machine learning on iOS
- 4:33Federated learning concept
- 12:38iOS implementation
- 20:49Model preparation
- 23:39Training in Swift
- 26:47Weight extraction and replacement
- 31:24Experiment and results
- 36:27CoreML limitations
- 39:16Conclusion and outlook
- 40:20Q&A
related talks
48mFederated ML for Edge Applications
41mFederated Learning: Machine Learning on Decentralized Data
22mCore ML and Your App
12mFederated Learning në ML/AI - mësimi i decentralizuar
34mDecentralized Machine Learning
26mThe Blue Ocean of Machine Learning: Core ML
40mDeep Learning on iOS
38mFederated Learning and Its Application...