The Staircase Mechanism in Differential Privacy

By Quan Geng, Peter Kairouz, Sewoong Oh, Pramod Viswanath. 2015

174 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Sociology and Political Science

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Badih Ghazi
2. Pasin Manurangsi
3. Ravi Kumar
4. Rasmus Pagh
5. Quan Geng
6. Pramod Viswanath
7. Amer Sinha
8. Jun Zhao

In Sociology and Political Science

1. Quan Geng
2. Pramod Viswanath
3. Jun Zhao
4. Ravi Kumar
5. Rasmus Pagh
6. Badih Ghazi
7. Pasin Manurangsi
8. Xiaokui Xiao

Citations in the corpus

• Collecting and Analyzing Multidimensional Data with Local Differential Privacy (2019, methodology): “[21] propose Staircase mechanism, which achieves optimal performance for unbounded input values (e.”
• Differentially Private Aggregation in the Shuffle Model: Almost Central Accuracy in Almost a Single Message (2021, background): “We also point out that in the low privacy regime ( " 1 ), the staircase mechanism is known to significantly improve upon the Laplace mechanism (Geng & Viswanath, 2016; Geng et al., 2015); the former achieves MSE that is exponentially small in ⇥ ( " ) while the latter has MSE O (1 / " 2 ) .”
• The Optimal Noise-Adding Mechanism in Differential Privacy (2012, methodology): “We would like to point it out that the staircase mechanism has been extended to the multidimensional setting in [53], and the optimal mechanism is also studied in [54] for approximate differential privacy under a similar optimization framework.”
• Mutual Information Optimally Local Private Discrete Distribution Estimation (2016, background): “We notice that there are some other local private mechanisms by adding noises (e.g. Laplace noises [9], two-sided geometric noises [12][11]) on bit maps, but their mutual information bounds or distribution estimation performances are dominated by their binary version: binary randomized response.”
• Local Differential Private Data Aggregation for Discrete Distribution Estimation (2019, background): “The popular mechanisms for achieving differential privacy in the centralized setting include Laplace mechanism [29], two-sid ed geometric mechanism [30], [31] that inject calibrated noises”