Data ownership is a competitive advantage and a very valuable resource. In the coming years, the AI front runners will capture most of the value unlocked by powerful machine learning models. Many machine learning projects, however, fail because not enough data is available.
In this talk, Daniel will show that collaborating on machine learning models is rational, even among competitors in an environment where data security and regulations are getting more and more important. One solution is federated machine learning that lets companies collaborate by distributing model training instead of sharing data with each other.
- Data is the most valuable asset in most companies, data is actively targeted by adversaries and data security sucks
- Federated machine learning enables secure and private collaborative machine learning without sharing data
- It’s rational to work together to build the most powerful machine learning models, even among competitors