Withdraw of paper accepted to AISTATS 2020 on differentially private top-k selection

We decided to withdraw our paper “Differentially Private Top-k Selection for Unknown Data Domain” accepted to AISTATS 2020. Unfortunately, we found an error with the main algorithm created, which invalidates the claims related to it. For this reason, we decided to retract the paper before creating a camera-ready version.

Nonetheless, we continue our research on the topic.

Finally, we'd like to thank LinkedIn engineers David Durfee and Ryan Rogers for invaluable discussions.