Federated Learning and Its Application...
summary
LiM is an Android malware classifier built with federated, semi‑supervised ensemble learning that reduces shared model dimensionality to five parameters so client updates are hard to reverse, and uses a cloud private dataset to dampen outlier updates. The system runs federated rounds with local models on phones, shows higher F1 and far fewer false positives than a centralized baseline, and withstands poisoning experiments even with 50% compromised clients while preventing straightforward inference of installed apps.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 10:04Android security limits
- 16:33Federated learning basics
- 20:05Privacy and security challenges
- 27:32LIM model implementation
- 31:05Performance and security evaluation
- 35:20Future challenges and applications
- 37:03Conclusion and Q&A
related talks
41mFederated Learning: Machine Learning on Decentralized Data
42mFederated Learning: Unlocking the Values of Discrete Data for Enterprise
42mLearning Together, Distributed: An Introduction to Federated Learning
48mFederated ML for Edge Applications
22mCollaborative Self-Learning for Security, Fun and Profit
34mDecentralized Machine Learning
44mFederated ML on iOS: Protecting User Data in the Age of AI
26mMachine Learning without the Data #federatedlearning