Federated Learning to Protect Sensitive Health Data

Category: Insurance & Risk Transfer • Article #14 • Reading time: 5 minutes

Introduction

Health data sensitivity prevents sharing yet industry-wide models require diverse data. Federated learning enables model improvement without centralizing sensitive health records.

Consortium Approach and Privacy

Insurers collaborate through federated learning improving models while protecting individual insurer data.

Results and Improvements

Federated health models improve accuracy 10-15% compared to individual insurer models.

Conclusion

Federated learning enables collaboration while protecting health data privacy.