On Privacy Protecting
#6
by
LarryHawkingYoung
- opened
It is mentioned in the dataset description that federated learning can be used to protect each user's privacy, so, is privacy preserving a requirement in our solution? If so, How dose the judgement platform evaluate whether our solution has protected privacy well or not? I can see from the leaderboard that only the prediction accuracy is considered.
The federated learning solution is encouraged to be used on this dataset, as the data is perfectly suitable for FL. However, the challenge results is only based on the prediction performance of the final model, whether it is an FL solution or a centralized one.
mmkamani7
changed discussion status to
closed