Industrial AI Podcast KI und federated learning im Spritzgießen
Sep 22, 2021
Michael Kuehne-Schlinkert, Founder and CEO of Katulu, delves into federated learning for machine building. He explains how this innovative technology allows machines to learn from one another while maintaining data privacy. The discussion highlights optimizing injection molding processes and the essential number of machines needed for effective model training. Kuehne-Schlinkert emphasizes the collaboration required among manufacturers to harness federated learning's full potential and create a robust European automotive ecosystem.
AI Snips
Chapters
Transcript
Episode notes
Federated Learning Explained Simply
- Federated Learning enables machines to learn from each other without sharing raw data.
- This concept leverages local model training and collective improvement without violating data privacy.
Evaluate Federated Learning Fit
- Evaluate if federated learning fits your company's digital strategy before implementing.
- Use pilot projects to test fits and validate use cases thoughtfully to ensure success.
Local Model Adaptation Advantage
- Federated Learning allows adapting AI models to specific machines and conditions without centralized data.
- This lets domain knowledge be applied uniquely per machine through local training.