1.4 What's the advantage of using Federated Validation?

Privacy and personal data are important. In most cases, it's not expected to obtain the datasets and use them outside of the institutes.

How to validate the a trained AI model becomes an issue. Without accessing the data insides other institutes, it's hard to assert the AI model performance; also, the AI model might need to get validated in more than one places.

Federated Validation system is designed to fix this dilemma. The AI model is kept on the owner, while data are kept in original place. With a simple click, the AI model will download automatically to the edge machine to get validated by local data. After the validation is complete, the AI model will be removed and the validation result will be sent to the FV aggregator to generate the overall validation reports. For more information about how this FV system works, please refer to the User Manuals.