Secure FLOATING - Scalable Federated Learning Framework for Real-time Trust in...
summary
Secure-FLOATING combines federated learning with secure multi-party computation (SMPC) and blockchain to validate trajectory data in real time without central servers. Nodes locally predict a neighbor’s next-10s trajectory with lightweight RNN/LSTM models, exchange secret-shared loss updates via SMPC (reduced message rounds) for aggregation, and record majority-rule endorsements on a permissioned blockchain/IPFS ledger. Evaluated on realistic NYC traces up to 8,000 nodes, it reports lower delay and overhead and up to 75% successful endorsement with 50% attacker penetration.
AI-written summary · about
chapters (10)
- 0:00Secure FLOATING project intro
- 1:56Centralized trust challenges
- 3:41Federated learning for privacy
- 6:13Blockchain consensus for trajectories
- 8:22Fed-Verified with SMC
- 10:50Evaluation setup and tools
- 11:52Lightweight trajectory models
- 12:55Resilience against malicious nodes
- 13:19Current challenges and future work
- 13:42Conclusion and collaboration
related talks
27mCreate Value and Insight Through Privacy-preserving Data Collaboration [...]
25mData collaboration with Federated Learning
06mPMF: A Privacy-preserving Human Mobility Prediction Framework via Federated Learning
42mFederated Learning: Unlocking the Values of Discrete Data for Enterprise
28mA Privacy Enhancing Federated Learning Framework for Multi-Party Distributed ..
31mFrom Theory to Practice: Advances in Real-World Federated Learning
16mPetalGuard, A Privacy-preserving Federated Learning Framework
42mLearning Together, Distributed: An Introduction to Federated Learning