Vertical Federated Learning across the Value Chain
summary
Katulu applies vertical federated learning across industrial value chains by slicing models into a U-shape split model and having suppliers train local layers while a neutral coordinator aggregates encrypted activations and performs private set intersections for batch IDs. Inference runs locally at the active manufacturer inside an encapsulated client to prevent reverse engineering. The system ties disparate supplier features to end-product labels so manufacturers adjust process settings to avoid over‑spec production and use supplier data without sharing raw data.
AI-written summary · about
chapters (5)
- 0:00Start
- 4:36Vertical FL Configurations
- 6:19The right Quality along the Value Chain
- 8:48Vertical FL along the Value Chain
- 11:26Data in the Value Chain
related talks
42mFederated Learning: Unlocking the Values of Discrete Data for Enterprise
27mCreate Value and Insight Through Privacy-preserving Data Collaboration [...]
48mFederated ML for Edge Applications
25mData collaboration with Federated Learning
41mFederated Learning: Machine Learning on Decentralized Data
29mBalancing Thin Line Between Data Intelligence & Privacy
30mFederated Learning from simulation to production
38mFederated Learning and Its Application...